Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

The use of public sector data analytics in the Netherlands

The use of public sector data analytics in the Netherlands Purpose – Public sector data analytics concerns the process of retrieving data, data analysis, publication of the results as well as re-using the data by government organizations to improve their operations and enhance public policy. This paper aims to explore the use of public sector data analytics in the Netherlands and the opportunities and challenges of this use. Design/methodology/approach – This paper finds 74 applications of public sector data analytics, identified by a Web search and consultation with policymakers. The applications are categorized by application type, organization(s) involved and application domain, and illustrative examples are used to elaborate opportunities and challenges. Findings – Public sector data analytics is most frequently used for inspection and enforcement of social services and for criminal investigation. Even though its usage is often experimental, it raises concerns for scope creep, repeated targeting of the same (group of) individuals, personal data use by third parties and the transparency of governmental processes. Research limitations/implications – Drawing on desk research, it was not always possible to identify which type of data or which technology was used in the applications that were found. Furthermore, the case studies are illustrative rather than providing an in-depth overview of opportunities and challenges of the use of data analytics in government. Originality/value – Most studies either perform a literature overview or present a single case study; this paper presents a more comprehensive overview of how a public sector uses data analytics. Keywords Data analytics, Big data, Open data, Artificial intelligence, Evidence-based policy-making, Data-driven policy-making Paper type Research paper © Anne Fleur van Veenstra, Francisca Grommé and Somayeh Djafari. Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) licence. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and non-commercial purposes), subject to full attribution to the original publication Transforming Government: People, Process and Policy and authors. The full terms of this licence may be seen at http://creativecommons.org/licences/by/4.0/ legalcode pp. 396-419 Emerald Publishing Limited This paper is based on the study “Quickscan AI in de publieke dienstverlening,” which was 1750-6166 DOI 10.1108/TG-09-2019-0095 commissioned by the Ministry of the Interior and Kingdom Relations of the Netherlands. 1. Introduction The use of Following developments in the private sector, governments around the world use data public sector analytics to capture the value of data and improve their way of working (Kim et al.,2014). data analytics Public sector data analytics concerns the data retrieval process, analysis of data, publishing the results as well as re-using the data to address societal challenges (Klievink et al.,2017; Lnenicka and Komarkova, 2019; Ooijen et al.,2019). It is driven by the increased availability of open and big data and data processing techniques (Janssen and Kuk, 2016; Van Veenstra and Kotterink, 2017). Furthermore, attention to data-driven methodologies is growing as governments aim to improve the effectiveness of their policies by making them more evidence-based (Höchtl et al.,2016; Giest, 2017; Poel et al., 2018). The use of public sector data analytics, however, has also led to concerns for the privacy of individuals and the transparency of government operations and decisions (Janssen and van den Hoven, 2015). Studies of public sector data analytics in the literature on digital government often focus on data availability for public or economic value (Susha et al.,2015; Millard, 2018). Actual usage has, so far, attracted much less attention and empirical evidence on which value is created or which concerns prove to be grounded is fragmented (Jagadish, 2015; Janssen and Kuk, 2016). Not surprisingly, few comprehensive empirical studies of data analytics in the public sector have been carried out; instead, research often presents a literature overview (cf. Mehr, 2017; Wirtz et al., 2019; Lnenicka and Komarkova, 2019) or individual case studies (cf. Zheng et al., 2018; Androutsopoulou et al., 2019; Sun and Medaglia, 2019). Therefore, we aim to contribute to closing this gap in the literature by investigating usage in the Netherlands based on a large number of applications. By analyzing current use of public sector data analytics including opportunities and challenges, we substantiate the relevance of these opportunities and challenges. A study of the Netherlands, a digital frontrunner within Europe (cf. its third ranking in the Digital Economy and Society Index [1]), is expected to yield state-of-the-art insights. Although the public sector has used data analytics for decades, we aim to investigate current developments and, therefore, focus on current usage and usage that ended in 2015 or later (also see the method section regarding selection and identification). Based on a Web search, we identify and analyze 74 applications in the public sector (see Appendix). We then further elaborate on the main findings using illustrative examples from Dutch practice. Before presenting these data, the next section provides a review of previous literature on the topic. We end the paper with a discussion, followed by conclusions and recommendations for further research. 2. Public sector data analytics To investigate the use of public sector data analytics, we first look at its drivers, followed by an overview of different types of use, and opportunities and challenges. The use of public sector data analytics is driven on the one hand by technological developments, such as the increased availability of data for re-use and the advance of data processing technologies. Data availability is increased by the publication and re-use of governmental data as open data (Janssen et al., 2012) and by the combination of a variety of data sets from different organizations (including the private sector) as big data (McAfee and Brynjolfsson, 2012; Manyika et al., 2011). Data processing technologies includes artificial intelligence (AI), which refers to systems that solve problems, take decisions or carry out tasks that require intelligence (Russell and Norvig, 2010) enable government organizations to capture the value of the available data. On the other hand, public sector data analytics is driven by government organizations aiming to address societal issues and inform the different steps of policy-making by making it more evidence-based (Höchtl et al.,2016; Giest, 2017). Following these two drivers also makes it possible to discern different types of use. TG Several studies identified (technical) applications of data analytics (Mehr, 2017; Chui et al., 15,4 2018; Wirtz et al., 2019). Mehr (2017), for example, categorizes the types of (technical) government problems appropriate for the use of AI, which include resource allocation, large data sets, a shortage of expertise, predictable scenarios and procedural and diverse data. Based on an analysis of advanced usage in different industries, Chui et al. (2018) find three main categories to improve performance: predictive maintenance, logistics optimization and personalization. And a list of ten potential technical applications of analytics in the public sector is provided by Wirtz et al. (2019), ranging from process automation to speech analytics. These types of usage focus on technical possibilities, but do not link to specific roles of government organizations. Others have identified types of usage that aim to improve policy-making (Santiso and Roseth, 2018; Poel et al., 2018; Ooijen et al., 2019). Santiso and Roseth (2018) and Ooijen et al. (2019) distinguish four types of analytics – descriptive, diagnostic, predictive and prescriptive – for different steps of policy-making. Poel et al. (2018) also distinguish different phases of policy-making, including foresight and agenda setting, monitoring and interim evaluation, problem analysis, identification and design of policy options, policy implementation and ex post evaluation and impact assessment. Furthermore, other types of usage are found by looking at individual illustrations of its use in public service delivery (Zheng et al.,2018; Androutsopoulou et al.,2019). Few of these studies have focused on actual usage, and none have looked at which types of governmental organizations use data analytics, in which domain, and which societal challenges are addressed [with the exception of Poel et al. (2018) that listed different policy areas]. Opportunities that are associated with the use of public sector data analytics are making governments’ operations more efficient and effective (Janssen and Kuk, 2016; Klievink et al., 2017) and increasing transparency, which is linked to openness and accountability of governmental processes (Meijer, 2009; Matheus and Janssen, 2015). Furthermore, opportunities for the use of data analytics for policy making are anticipatory governance, design and delivery and performance management (Ooijen et al.,2019). There are, however, also challenges associated with the use of public sector data analytics. Technical challenges are ensuring data quality and security (Bertot and Choi, 2013; Janssen et al.,2012). Also the re-use of data requires the presence of a data infrastructure ensuring interoperability (Lnenicka and Komarkova, 2019; Ooijen et al.,2019), and attention should be given to linking data semantically (Bizer, 2009), to allow for findability and re-use. Besides linking data technically, many studies have found that merely publishing data is insufficient to enable and stimulate re-use and, therefore, call for supporting collaborations with relevant stakeholders (Susha et al.,2017; Styrin et al.,2017) as well as for co-creation with citizens (Ferro et al.,2013; Charalabidis et al.,2014; van Veenstra and Kotterink, 2017). Challenges associated with the use of data analytics for policy-making are the public sector’s limited organizational readiness (Klievink et al., 2017), including governments’ hesitance to rely on new or emerging data sources (Poel et al.,2018), decision-makers’ unwillingness to be informed by data science (van der Voort et al.,2019) and negative consequences for operational employees (Barocas and Selbst, 2016). Furthermore, research shows that the expectation of many government organizations to become transparent as a result of using data analytics too is often high (Bannister and Connolly, 2012). The emergence of bias as a result of the use of algorithms is also considered a major challenge (Janssen and Kuk, 2016). Bias among others originates in the training data used for developing and refining algorithms, and can eventually affect decision-making (Hacker, 2018). Finally, legal concerns include privacy and accountability of governmental decision- making, for instance, when the application of a technology changes scope (“scope creep”) The use of and ceases to be proportional (Janssen and van den Hoven, 2015; Lourenço et al.,2015; public sector Zoonen, 2016, 2019; Agarwal, 2018). data analytics This literature review makes clear that public sector data analytics is driven by the application of innovations that have the potential to improve policy-making and address societal challenges, but many technical as well as organizational and legal challenges exist and it is not yet clear how these opportunities and challenges are related to actual usage. 3. Method The investigation of public sector data analytics in the Netherlands took place between November 2018 and March 2019. For the identification of applications of public sector data analytics, a structured Web search resulted in 74 examples, in which we included different government levels (municipal, provincial and national) as well as public and semi-public agencies. The Web search covered (academic) search engines and key government databases [2]; these were searched using the terms “AI,”“AI AND government,”“Data AND government” and “Algorithm AND government.” Google, Google Scholar and PiCarta (the national library catalogue) were used for orientation and complementary information. Another source for orientation was LexisNexis, a research database covering a large range of print media in the Netherlands. For up-to-date information on public usage, we relied on the following government and semi-government databases: Overheid.nl (the main portal for government information) and the Pilot Starter database, which lists pilot projects conducted by municipalities. Overheid.nl is a central access point to all government information. We therefore searched this database for additional terms in full-text: “kunstmatige intelligentie” (AI), “experiment” and “pilot.” Using this combination of sources, several characteristics were captured and verified for each application, including application type, organization type, status (implemented or experimental) and domain of usage. These characteristics were chosen because they capture the nature and aims of usage (application type) or because they have not yet been structurally investigated (type of organization and domain of usage). The categorization of applications was performed by two researchers independently; when differences between the categorizations occurred, a discussion followed based on which a decision was made. The Web search results were reviewed by a panel of policymakers from different levels of government. This panel was formed by two representatives from the Ministry of the Interior, two from the Association of Dutch Municipalities and six from different municipalities (all involved in using data analytics). The review took place as a session during a workshop and was facilitated by one author of this paper, while another author took notes. The purpose of the session was to validate whether the applications represent types of use in the Netherlands. This panel also provided information on the current status of some applications and provided internal policy documents that helped to specify the types of data and analytics that were used (the latter details are often difficult to verify through a Web search). Finally, we acknowledge that the resulting overview cannot provide a comprehensive record of all instances of public sector data analytics. First, we needed to balance the identification of applications with collecting and verifying the characteristics of each project. Furthermore, not all applications are made public online and we expect our search terms to miss a number of initiatives. In addition, public sector data analytics is a dynamic field which limits the extent to which any list can capture all usage. Not knowing which applications we have missed, it is not possible to determine how representative this list is. Nevertheless, the chosen government databases and additional sources do provide a basis for identifying key types of usage and characteristics. We use illustrative examples (based TG on information from public documents, such as policy documents and newspaper articles) to 15,4 elaborate on the main findings on these types of usage and characteristics (see Section 5). 4. Analysis of the applications Based on the list of 74 applications, three analyses were undertaken: a categorization of the examples by application type and status; by type of organization and application type; and by application domain and type. 4.1 Types of applications To the best of our knowledge, there is no established categorization of public sector data analytics that captures types of use. Our literature review found technological and policy- making oriented categorizations, but none focusing on of usage. Therefore, we categorized the 74 applications into 6 types of usage, inspired by the literature overview and refined based on the applications found online: (1) Personalization: the tailoring of public service delivery to individual needs. (2) Resource allocation: optimizing logistics for efficient use of resources. (3) Maintenance: identifying potential risks to schedule repairs. (4) Inspection and enforcement: predicting (behavioral) patterns to prevent misuse of public spending. (5) Crime investigation: pattern recognition and prediction such as fraud. (6) Forecasting: predicting macro-economic trends and variables. These types of usage are a mix of operational issues that government organizations may wish to address in different phases of the policy-making process (problem definition, execution and monitoring) as well as for other roles government organizations (operations and service delivery) have. We found experimental applications, meaning that they are not yet part of the standard operational activities of government, as well as implemented practices that are fully operational and part of the standard activities of government organizations. Regarding the categorization shown in Figure 1, the category with the largest number of data analytics applications in our list of examples is “inspection and enforcement” (20 out of Figure 1. The application domains per type of application 74). This category also includes the largest group of implemented applications (8 out of 20). The use of A total of 17 out of 74 are categorized as “investigation”; 4 out of these 17 were found to be public sector implemented. A total of 11 out of 74 applications fit the category of “resource allocation” and data analytics 9 applications are categorized as “maintenance.” For the latter two categories, only one implemented application was found (for “resource allocation”). Of the 74 applications, 14 are categorized as “personalization”; 2 out of these 14 were implemented. Finally, the category for which we found the lowest number of applications is “forecasting” (3 out of 74); all 3 are experiments. Although the number of applications for each category can be an effect of the search strategy and availability, the findings are indicative of the important role of experimentation in all categories. It also suggests that inspection and enforcement, investigation and maintenance and resource allocation are active fields for experimentation and implementation. 4.2 Types of government organizations The types of government organizations (municipality, central government, province or water board [regional], collaboration, executive or semi-public organization or other) involved in the use of data analytics are presented in Figure 2. Of the 74 applications, 18 we found are collaborations between different levels of government or between public and semi-public or non-governmental organizations, such as companies and non-governmental organizations. Another 16 applications are collaborations between municipalities or regional governments and are categorized as municipal or regional. Apart from forecasting, we found a few (2–5 out of 18) applications for every application type within the identified collaborations, indicating collaborations are formed for various types of applications. We found municipalities and executive and semi-public organizations to each be involved in 24 of the 74 applications, while 5 applications were found for “other,” 2 applications for the national and 1 for the regional level. Although national and regional governments are also represented in several collaborations and the exact proportions may not reflect all current activity, these findings do suggest the active role of municipalities in public sector data analytics in the Netherlands. The applications we found on the municipal level most often concern “inspection and enforcement” (8 out of 25), “personalization” (5 out of 25) and “investigation” and “resource allocation” (both 4 out of 25), while we found fewer applications for “maintenance” and Figure 2. Types of government organizations per type of application “forecasting.” Within the group of applications of executive and semi-public organizations, 9 TG out of the 24 applications are categorized as “investigation” and 5 out of the 24 as 15,4 “inspection and enforcement.” Only two or three applications were found for the other categories. The latter cover all types except for forecasting, for which no applications were found. 4.3 Application domain To investigate which sectors use data analytics, finally, we focus on their application domain (Figure 3). Most applications were found in the physical environment, which includes the public sphere, water management, mobility and infrastructure (22 out of 74). In this domain, all types of applications of data analytics were found. Most applications were found in the categories “maintenance” (7 out of 22), “personalization” (5 out of 22) and “resource allocation” (5 out of 22). In the physical domain, usage can thus be considered to comprise all application types. Two other domains in which many applications were found are safety and security (20 out of 74) and social services (17 out of 74). Within the safety and security domain, the applications we found are overwhelmingly categorized as “investigation” (15 out of 20); in the social domain, a majority of applications concerned “inspection and enforcement” (10 out of 17). Compared to the physical domain, data analytics usage in these domains is more focused. For the other domains (data infrastructure and finance, which includes applications linked to the governmental data infrastructure and finance, economy, labor market and education; and other, comprising applications from the legal domain and in education), few applications were found and with various application types. Therefore, no uniform conclusions can be drawn for these domains. 5. Key findings, opportunities and challenges of data analytics In this section, we discuss four main findings based on the overview in Section 4. Our aim is to relate these findings to opportunities and challenges of public sector data analytics. The opportunities and challenges discussed in the context of each finding serve to indicate that research focused on gaining an overview of the field can also contribute to insights into how opportunities and challenges of data analytics emerge. These brief discussions are not Figure 3. The types of application and their implementation meant to be comprehensive, instead, we highlight one or two opportunities and challenges in The use of relation to each finding on the basis of prominent themes in the literature (Section 2). As public sector noted in the literature review, among the opportunities are increasing efficiency, data analytics effectiveness and transparency. Transparency also is a relevant challenge related to data analytics, as are legal concerns (privacy and accountability), bias and organizational readiness. A disadvantage of an overview study is that it does not support a more detailed understanding of types of projects that have been executed and how they develop in practice. This section therefore also presents illustrative examples of applications from the list of 74, drawing on publicly available documents. Without claiming to be generalizable, these examples (sensors in municipal waste containers; data mining for social welfare fraud; a chatbot; and data mining for crime investigation) were chosen because they represent typical uses of data analytics in the Dutch public sector, and we therefore expect them to be informative for our understanding of future developments. 5.1 Many applications are used for operational purposes Many applications found are for operational usage, such as resource allocation and inspection and enforcement. Forecasting was the least common usage we found. The opportunity of using data analytics for everyday operations of public services is exemplified by the placement of sensors in household waste containers by the municipality of Rotterdam to improve resource allocation and maintenance (Gemeente Rotterdam, 2020). A total of 6,500 sensors were used to compute the most efficient collection route, thereby not only attempting to increase the quality of services but also to decrease CO emissions. Furthermore, the sensors also keep track of container defects. After a pilot in 2017, the system has been scaled to the city level in 2018 (Gemeente Rotterdam, 2017). A challenge characteristic for this type of usage is the collection of data about mundane aspects of citizens’ lives by third parties that carry out activities for public service delivery. This is exemplified by the Rotterdam waste service, which relies on third parties for data collection and processing. For other operational services, some critics have noted that third party involvement is not only a privacy challenge but can also affect the transparency and accountability of public services as it can be difficult for citizens to find and address the third parties involved (Van Teeffelen and Naafs, 2017). 5.2 Most often identified applications concern “inspection and enforcement” Inspection and enforcement is the most common usage category among the identified applications. Furthermore, half of these applications were found in the social domain. The use of data analytics is, thus, often seen as an opportunity to efficiently prevent misuse of government funds and decrease social service fraud. Systeem Risico Indicatie (SyRI) or “risk indication system” is an example of the promise of data analytics to this end. Discontinued in 2020, SyRI was a data mining application combining a variety of data types, such as data on taxes, properties, education, pension, debt, benefits and health insurance to search for anomalies and create risk profiles (Olsthoorn, 2016). The system was used from 2014 by varying constellations of government organizations, including municipalities, the Inland Revenue and the Inspection of the Ministry for Social Affairs. An important concern relating to inspection and enforcement is its reliance on the analysis of large amounts of sensitive data. A risk is that such data can be collected and analyzed for large groups of people without specific grounds for suspicion. This risk is related to a starting point of many data mining methods: a more open-ended analysis of TG large amounts of data to find previously unknown behavioral patterns (Olsthoorn, 2016). In 15,4 the case of SyRI, civil rights advocates started a lawsuit against the state because of such practices, which they referred to as “fishing expeditions” (Platform Bescherming Burgerrechten, 2018). In February 2020, court agreed that such data usage is not proportional and ruled that SyRI was in breach of Article 8 of the EU Charter of Human Rights (the right to privacy). On top of this, court ruled that the SyRI program lacked transparency and accountability, as it did not provide access to the algorithm that generated risk profiles. A second concern relevant to highlight regarding inspection and enforcement is that data analytics practices may display a bias toward vulnerable groups. This also characterizes the social domain, as governments often collect more data about these groups (Barocas and Selbst, 2016; Janssen and Kuk, 2016). SyRI, for instance, predominantly targeted people in low-income neighborhoods (about whom more data are available). 5.3 Many applications concern experimental usage Over half of the applications found are pilots or experiments (even though for many applications, the implementation status could not be confirmed). Experimentation provides degrees of freedom needed to apply and develop a technology and explore its capacities. For instance, a chatbot piloted by the national service center for entrepreneurs (Rijksdienst voor Ondernemend Nederland, [RVO]) was used as a search engine for its website using natural language processing. The aim of this 2017 experiment was to facilitate entrepreneur compliance with complex legislation regarding agricultural manure policies (relevant to nitrogen emissions). The chatbot prototype was further developed and integrated into Facebook Messenger to redirect questions away from the call center (RVO, 2017). While personalization offers opportunities to improve service delivery toward citizens, to tailor (often complex) government information to users and to cut costs on service delivery, a key challenge is to ensure accessibility because it involves direct interaction between information technology systems and citizens. The RVO chatbot is a good illustration of the problems that can emerge when this is not successful. During the development of the first version of the system, the RVO testers commented that the chatbot came across as “robotic,” thereby lacking human nuance (RVO, 2018). Furthermore, call center employees were not able to ensure accessible service delivery at first because they could not assume control of the system. Another opportunity of experimentation is to enroll stakeholders without demanding long-term commitment. FinPro is an example of criminal investigation that took advantage of the experimental format’s flexibility. The name refers to a collaboration established for data experiments by platform organization MeetingMoreMinds in 2011, which included, among others, the Rotterdam police, the Chambers of Commerce, health insurance companies, travel organizations and bus companies (the collaboration has not formally been disbanded, although most activities were conducted before 2017). The main purpose was to identify behavioral patterns on an aggregated level. The experimental format allowed for trialing the ad hoc involvement of experts from different domains to analyze data and interpret findings; a feature that initially led to new insights (Roobeek, 2016). However, the use of experimental data analytics may entail transparency issues as not all pilots are well documented and made available to public scrutiny. 5.4 Many data analytics projects are executed by multi-stakeholder collaborations The use of Nearly half of the identified applications concern collaborations: between the same type of public sector organizations, between different levels of government or between public and semi-public or data analytics non-governmental organizations. Collaborations allow for combining data sets to identify new patterns or anomalies in the data, and for attracting financial and technical resources, as illustrated by FinPro (Roobeek, 2016). Yet, a challenge of collaboration is “scope creep”: when the application of a technology broadens its scope beyond its original intensions. This also means that legal concerns arise that were not identified in advance. For instance, the FinPro collaborators learned along the way that the outcomes of data analysis can often not be submitted as legal evidence (Jessayan and De Lange, 2016). To return to the example of SyRI discussed above, enlarging the number of involved stakeholders entailed multiplying the number of interests and aims invested in the project. In SyRI, the scope was broadened over the years from welfare fraud to include other social issues such as harassment and nuisance (Raad van State, 2014). Scope creep thus raises legal concerns as the purpose of data collection and analysis may not be proportionate to the infringement of personal sphere. To conclude, we elaborated on four main findings to indicate that research focused on improving our overview of the field can also contribute to insights to how opportunities and challenges of data analytics emerge; and support more detailed insights into types of projects that have been executed and how they develop in practice. We further discuss our findings in the next section. 6. Discussion Our four main findings both corroborate and contribute to current research on public sector data analytics. First, we found that many applications aim for efficiency and effectiveness of government operations rather than for increasing transparency or facilitating evidence-based policy-making. This corroborates findings from Giest (2017) and Poel et al. (2018) that the use of data analytics for policy is still in its early days. Furthermore, third parties’ data analytics activities in operational processes of the public sector lead to concerns of third party access to data about everyday citizen behavior and of transparency and accountability of these processes. These concerns have been amply addressed in the literature about smart cities (Zoonen, 2016) but are equally relevant to research on digital government. Second, a frequent use type among the applications we identified is inspection and enforcement for social services (often used by municipalities). This implies that, within operational applications, data analytics is often used for investigation or enforcement of legislation. This makes the societal, ethical and legal concerns connected to this use type key in studies of public sector data analytics [see, for example, Janssen and van den Hoven (2015)]. Next to proportionality and invasiveness, the repeated targeting of (groups of) individuals based on their socio-economic class (or that of their neighborhood) also emerged as a concern. We know from the wider literature on the risk of bias in big data usage that when targeted more often, citizens are also at greater risk of becoming “anomalies in the data” and objects of further investigation (Barocas and Selbst, 2016), with an associated risk of stigmatization. Third, as more than half of the identified applications are experiments, this implies that, as suggested by the literature, many applications of data analytics face difficulties moving beyond the experimentation phase because of technical and legal challenges [see, for instance, van Veenstra and Kotterink (2017)]. Similarly, personalization and tailoring of services still seem to be under development, implying that data analytics are neither easily deployed effectively for this purpose nor lead to more accessibility in the experience of TG citizens. As emphasized in the literature (Klievink et al.,2017), to gain from the opportunities 15,4 that these technologies may offer, attention also needs to be given to organizational readiness and change. As well as experimental, many projects are collaborative, which is our fourth finding. From two of our illustrative examples, we furthermore learn that these collaborations can involve changing stakeholder constellations. Operating in a multi- actor context, as is the case for many data analytics initiatives in the Netherlands, is an additional challenge for guarding the scope of projects. Furthermore, this coincides with an observation regarding the applications we identified: we find that much attention is given to data sharing among different stakeholders rather than to using more advanced processing technologies such as AI (Sun and Medaglia, 2019); applications of collaborative data analytics often combine data sets and search for behavioral patterns or anomalies. While this study was one of a very few providing an overview of public sector data analytics based on a larger number of cases, a main limitation is its reliance on desk research. Therefore, it was often not possible to present details such as the type of analytics that was used. Consequently, it is impossible to draw conclusions based on this study about which specific technologies are used by government. A first avenue for further research, therefore, is to examine which technologies are used. A second avenue is examining the opportunities and challenges of public sector data analytics more structurally and in-depth, capturing also the views of those involved. 7. Conclusion This study explores the use of public sector data analytics in the Netherlands and aims to contextualize opportunities and challenges identified in previous research. Based on a Web search, we find that many identified applications are used for operational processes, and are often deployed in processes of inspection and enforcement for social services. Furthermore, many applications are pilots or experiments rather than routine usage and they concern collaborations between different stakeholders. We thus find that advance data processing is still in its early stages. At the same time, the illustrative examples raise several concerns including third-party access to personal data, repeated targeting of the same (groups of) individuals, limitations to accessibility and transparency and scope creep, especially in multi-actor settings. This implies that the use of public sector data analytics requires developing organizational capabilities to ensure effective use, foster collaboration and scale up, as well as legal and ethical capabilities to carefully balance these concerns with their objectives, such as increased operational efficiency or effectiveness. As a major limitation of this study is its reliance on desk research, further research should perform more in-depth research structurally capturing the opportunities and challenges of the use of the application types of public sector data analytics. Notes 1. The Digital Economy and Society Index is a composite index with digital public services as one of the components. For more information, see: https://ec.europa.eu/digital-single-market/en/desi 2. The source URLs are: google.com; scholar.google.com; picarta.nl; lexisnexis.nl; overheid.nl; and depilotstarter.vng.nl. References The use of Agarwal, P.K. (2018), “Public administration challenges in the world of AI and bots”, Public public sector Administration Review, Vol. 78 No. 6, pp. 917-921. data analytics Androutsopoulou, A., Karacapilidis, N., Loukis, E.N. and Charalabidis, Y. (2019), “Transforming the communication between citizens and government through AI-guided chatbots”, Government Information Quarterly, Vol. 36 No. 2, pp. 167-384. Bannister, F. and Connolly, R. (2012), “The trouble with transparency: a critical review of openness in e- Government”, Policy and Internet, Vol. 3 No. 1, pp. 1-30. Barocas, S. and Selbst, A.D. (2016), “Big data’s disparate impact”, California Law Review, No. 104, pp. 671-732. Bertot, J.C. and Choi, H. (2013), “Big data and E-government: issues, policies, and recommendations”,in Proceedings of the 14th Annual International Conference on Digital Government Research. dg.o. 2013, 17–20 June, Quebec City, Canada, pp. 1-10. Bizer, C. (2009), “The emerging web of linked data”, IEEE Intelligent Systems, Vol. 24 No. 5, pp. 87-92. Charalabidis, Y., Loukis, E.N., Androutsopoulou, A., Karkaletsis, A. and Triantafillou, A. (2014), “Passive crowdsourcing in government using social media”, Transforming Government: People, Process and Policy, Vol. 8 No. 2, pp. 283-308. Chui, M., et al. (2018), Notes from the AI Frontier: Applications and the Value of Deep Learning Insights from Hundreds of Cases, McKinsey Global Institute, available at: www.mckinsey.com//media/ McKinsey/Featured%20Insights/Artificial%20Intelligence/Notes%20from%20the%20AI%20 frontier%20Applications%20and%20value%20of%20deep%20learning/Notes-from-the-AI- frontier-Insights-from-hundreds-of-use-cases-Discussion-paper.ashx Ferro, E., Loukis, E.N., Charalabidis, Y. and Osella, M. (2013), “Policy making 2.0: from theory to practice”, Government Information Quarterly, Vol. 30 No. 4, pp. 359-368. Gemeente Rotterdam (2017), “Persbericht: Minder kosten en CO2 uitstoot door slimme afvalcontainers”, Rotterdam.nl, available at: www.persberichtenrotterdam.nl/bericht/2270/Minder-kosten-en-CO2- uitstoot-door-slimme-afvalcontainers/ Gemeente Rotterdam (2020), “Slimme wijkcontainers”, Rotterdam.nl, available at: www.rotterdam.nl/ wonen-leven/slimme-wijkcontainers/ Giest, S. (2017), “Big data for policymaking: fad or fasttrack?”, Policy Sciences, Vol. 50 No. 3, pp. 367-382. Hacker, P. (2018), “Teaching fairness to artificial intelligence: Existing and novel strategies against algorithmic discrimination under EU law”, Common Market Law Review, Vol. 55 No. 4, pp. 1134-1185. Höchtl, J., Parycek, P. and Schöllhammer, R. (2016), “Big data in the policy cycle: Policy decision making in the digital era”, Journal of Organizational Computing and Electronic Commerce, Vol. 26 Nos 1/2, pp. 147-169. Jagadish, H.V. (2015), “Big data and science: Myths and reality”, Big Data Research, Vol. 2 No. 2, pp. 49-52. Janssen, M. and Kuk, G. (2016), “The challenges and limits of big data algorithms in technocratic governance”, Government Information Quarterly, Vol. 33 No. 3, pp. 371-377. Janssen, M. and van den Hoven, J. (2015), “Big and open linked data (BOLD) in government: a challenge to transparency and privacy?”, Government Information Quarterly, Vol. 32 No. 4, pp. 363-368. Janssen, M., Charalabidis, Y. and Zuiderwijk, A. (2012), “Benefits, adoption barriers and myths of open data and open government”, Information Systems Management,Vol.29 No. 4, pp. 258-268. Jessayan, H. and De Lange, B. (2016), “Big data legt onzichtbare criminaliteit bloot”, Het Financieele Dagblad, 9 June. Kim, G.-H., Trimi, S. and Chung, J.-H. (2014), “Big data applications in the government sector”, TG Communications of the Acm, Vol. 57 No. 3, pp. 78-85. 15,4 Klievink, B.,Romijn,B.-J.,Cunningham,S., Bruijn,H.and de,(2017), “Bigdata inthe public sector: Uncertainties and readiness”, Information Systems Frontiers,Vol.19 No. 2, pp. 267-283. Lnenicka, M. and Komarkova, J. (2019), “Big and open linked data analytics ecosystem: Theoretical background and essential elements”, Government Information Quarterly, Vol. 36 No. 1, pp. 129-144. Lourenço, R.P., Piotrowski, S. and Ingrams, A. (2015), “Open data driven public accountability”, Transforming Government: People, Process and Policy, Vol. 11 No. 1, pp. 42-57. McAfee, A. and Brynjolfsson, E. (2012), “Big data: the management revolution”, Harvard Business Review, No. October, pp. 1-9. Manyika, J., Chui, M., Brown, B., Bughin, J., Dobbs, R., Roxburgh, C. and Hung Byers, A. (2011), Big Data: The Next Frontier for Innovation, Competition, and Productivity, McKinsey Global Institute, available at: www.mckinsey.com/business-functions/mckinsey-digital/our-insights/ big-data-the-next-frontier-for-innovation Matheus, R., Janssen, M., et al. (2015), “Transparency dimensions of big and open linked data transparency as being synonymous with accountability and openness”, in: Janssen (Eds.) I3e 2015, Lncs 9373, Springer, pp. 236-246. Mehr, H. (2017), “Artificial intelligence for citizen services and government”, Harvard Ash Center, available at: http://ash.harvard.edu/files/ash/files/artificial_intelligence_for_citizen_services. pdf, Meijer, A. (2009), “Understanding modern transparency”, International Review of Administrative Sciences, Vol. 75 No. 2, pp. 255-269. Millard, J. (2018), “Open governance systems: Doing more with more”, Government Information Quarterly, Vol. 35 No. 4, pp. 77-87. Olsthoorn, P. (2016), “Big data voor fraudebestrijding”, WRR Working paper no. 21, available at: www. wrr.nl/publicaties/working-papers/2016/04/28/big-data-voor-fraudebestrijding Ooijen, C., van Ubaldi, B. and Welby, B. (2019), “A data-driven public sector”, OECD Working Papers on Public Governance No. 33. Platform Bescherming Burgerrechten (2018), “NGO’s starten rechtszaak tegen de staat over risicoprofilering van nederlandse burgers”, December 1, available at: https://platform burgerrechten.nl/2018/01/12/ngos-starten-rechtszaak-tegen-de-staat-over-risicoprofilering-van- nederlandse-burgers/ Poel, M., Meyer, E.T. and Schroeder, R. (2018), “Big data for policymaking: Great expectations, but with limited progress?”, Policy and Internet, Vol. 10 No. 3, pp. 347-367. Raad van State (2014), “Advies W12.14.0102/III”, available at:www.raadvanstate.nl/adviezen/advies. html?id=11339 Roobeek, A. (2016), “Finpro geeft met verantwoorde Big-Datamethoden zicht op onzichtbare veiligheid”, Tijdschrift Voor de Politie, Vol. 78 No. 9, pp. 24-28. Russell, S.J. and Norvig, P. (2010), Artificial Intelligence: A Modern Approach, Third edition, Pearson Education, Upper Saddle River, NJ. RVO (2017), Eindrapport Pilot IBM Watson, The Hague, RVO. RVO (2018), Presentatie: Bestuurlijke Bijeenkomst Artificial Intelligence (AI)/Chatbot, The Hague, RVO. Santiso, C. and Roseth, B. (2018), “Data disrupts corruption”, Stanford Social Innovation Review, Spring, pp. 50-55. Styrin, E., Luna-Reyes, L.F. and Harrison, T.M. (2017), “Open data ecosystems: an international comparison”, Transforming Government: People, Process and Policy,Vol.11 No. 1, pp. 132-156. Sun, T.Q. and Medaglia, R. (2019), “Meeting the challenges of artificial intelligence in the public The use of sector: Evidence from public healthcare”, Government Information Quarterly, Vol. 36 No. 2, public sector pp. 368-383. data analytics Susha, I., Grönlund, A. and Janssen, M. (2015), “Organizational measures to stimulate user engagement with open data”, Transforming Government: People, Process and Policy, Vol. 9 No. 2, pp. 181-206. Susha, I., Janssen, M. and Verhulst, S. (2017), “Data collaboratives as “bazaars?”, Transforming Government: People, Process and Policy, Vol. 11 No. 1, pp. 157-172. van der Voort, H.G., Klievink, A.J., Arnaboldi, M. and Meijer, A.J. (2019), “Rationality and politics of algorithms. will the promise of big data survive the dynamics of public decision making?”, Government Information Quarterly, Vol. 36 No. 1, pp. 27-38. Van Teeffelen, K. and Naafs, S. (2017), “Moeten We wel zo blij zijn met de slimme stad?”, Trouw,5 December, available at: www.trouw.nl/samenleving/moeten-we-wel-zo-blij-zijn-met-de-slimme- stad-aaa1a488/ Van Veenstra, A.F., Kotterink, B., et al. (2017), “Data-Driven policy making: the policy lab approach”,in: Parycek (eds.) ePart 2017, LNCS 10429, Springer, pp. 100-111. Wirtz, B.W., Weyerer, J.C. and Geyer, C. (2019), “Artificial intelligence and the public sector – applications and challenges”, International Journal of Public Administration, Vol. 42 No. 7, pp. 596-615. Zheng, Y., Yu, H., Cui, L., Miao, C., Leung, C. and Yang, Q. (2018), “SmartHS: an AI platform for improving government service provision”, Association for the Advancement of Artificial Intelligence, January 22, available at: www.ntulily.org/wp-content/uploads/conference/Smart HS_An_AI_Platform_for_Improving_Government_Service_Provision_accepted.pdf Zoonen, L.V. (2016), “Privacy concerns in smart cities”, Government Information Quarterly, Vol. 33 No. 3, pp. 472-480. Zoonen, L.V. (2019), “Opnieuw fatale remedies”, Sociologie, Vol. 15 No. 1, pp. 19-43. Appendix TG 15,4 Table A1. Applications of public sector data analytics in The Netherlands Type of Application (name) Organizations involved Domain application Status Source URL SyRI UWV, SVB, Inspectie SZW, Social services Inspection and Implemented https://wetten.overheid.nl/BWBR0013060/ Belastingdienst and municipalities enforcement 2019-01-01 Criminaliteits Politie Safety/security Investigation Unknown www.politie.nl/nieuws/2017/mei/15/05-cas. Anticipatie Systeem html (CAS) Infobox crimineel en Openbaar Ministerie, Politie, Safety/security Investigation Implemented Stcrt. 2015, nr. 7509 and https://hetccv.nl/ onverklaarbaar Belastingdienst, FIOD and onderwerpen/veilige-vakantieparken/ vermogen (iCOV) Financial Intelligence Unit, overzicht-maatregelen/analyse-icov/; www. Inspectie SZW, Inlichtingen- en securitymanagement.nl/big-data-fraude/? Opsporingsdienst van de Inspectie vakmedianet-approve-cookies=1&_ga= voor de Leefomgeving en Transport 2.167311545.627848040.1548922282- (ILT-IOD), Inlichtingen- en 359266002.1548922282 Opsporingsdienst van de Nederlandse Voedsel- en Warenautoriteit (NVWA-IOD), Commissariaat voor de Media iRN_iColumbo Politie Safety/security Investigation Unknown TK 2013–2014, 26 643, nr. 298, p. 16 Business Intelligence Belastingdienst Government Inspection and Implemented https://decorrespondent.nl/1766/Vergeet-de- and Analytics (BI&A) finances enforcement politiestaat-Welkom-in-de-belastingstaat/ 90525160-5a2c27c2 Verwijzingsindex Financial Intelligence Unit Safety/security Investigation Implemented https://wetten.overheid.nl/BWBR0041193/ Recherche Nederland (FIU-Nederland) 2018-07-25 Onderzoeken en Subjecten (VROS- bestand) Parkeersensoren voor Municipality of Zoetermeer Physical Resource Experiment https://depilotstarter.vng.nl/projecten/ een dynamisch environment allocation omgevingswet/parkeersensoren-voor-een- verkeers- (mobility) dynamisch-verkeersverwijssysteem verwijssysteem (continued) The use of public sector data analytics Table A1. Type of Application (name) Organizations involved Domain application Status Source URL Waterproef/ Municipalities in Friesland, Social services Inspection and Unknown www.hoogeveen.nl/bis/dsresource? waterproof Groningen, Drenthe, SVB, enforcement objectid=d3525063-c977-4300-b159- Belastingdienst d114a3810a04 Project Regionale 130 municipalities Social services Inspection and Unknown TK 2010–2011, 17050, nr. 414 (bijlage) Coördinatiepunten enforcement Fraudebestrijding Noord, Overijssel, Gelderland, Flevoland Schoon schip Municipality of Amsterdam, UWV, Social services Inspection and Unknown TK 2010–2011, 17050, nr. 414 (bijlage) Belastingdienst, SVB enforcement E-Discovery Nationaal Archief, ICTU Data Maintenance Experiment www.nationaalarchief.nl/archiveren/ management nieuws/e-discovery-machine-learning-en-e- mail Kamer van Kamer van Koophandel Economy Inspection and Implemented TK 2010–2011, 17050, nr. 414 (bijlage) Koophandel enforcement DUO DUO Education Inspection and Implemented TK 2010–2011, 17050, nr. 414 (bijlage) enforcement WGA project Gemeente Vlissingen Social services Inspection and Unknown TK 2012–2013, 33 400-XV, nr. 5 Middengebied enforcement Vlissingen Rechtmatigheids- Inlichtingenbureau Social services Inspection and Implemented TK 2012–2013, 33 400-XV, nr. 5; www. controle enforcement inlichtingenbureau.nl/ Inlichtingenbureau Team Darkweb Financial Intelligence Unit Safety/security Investigation Experiment TK 2017–2018, 29 911, nr. 200, p.12 Nederland (FIU-Nederland) (criminal activity) Organized Crime Financial Intelligence Unit Nederland Safety/security Investigation Unknown TK 2017–2018, 29 911, nr. 200, p.11 Field Lab (FIU-Nederland), Nationale Politie, (criminal Openbaar Ministerie, municipalities activity) of Amsterdam and Rotterdam, Belastingdienst, FIOD, Inspectie SZW (continued) TG 15,4 Table A1. Type of Application (name) Organizations involved Domain application Status Source URL SIGMA Nationale Politie, Koninklijke Migration Inspection and Implemented TK 2017–2018, 29 911, nr. 200, p.13; www. Marechaussee, Expertisecentrum enforcement rijksbegroting.nl/2016/verantwoording/ Mensenhandel en Mensensmokkel, jaarverslag,kst233798_13.html þwww. IND, Inspectie SZW, Openbaar rijksictdashboard.nl/projecten/301498 Ministerie DigiDijk Rijkswaterstaat, Physical Maintenance Unknown https://zoek.officielebekendmakingen.nl/ Hoogheemraadschap Delfland environment blg-820189.pdf (water management) Landelijk Meetnet – Physical Inspection and Experiment www.overheid.nl Luchtkwaliteit environment enforcement (air quality) FinPro Openbaar Ministerie, Municipality Safety/security Investigation Experiment www.meetingmoreminds.com/wat-doen-we/ of Rotterdam, Programmabureau (criminal overige-projecten-initiatieven/ NPRZ, Politie activity) Debt Alert CJIB Government Personalization Experiment https://ibestuur.nl/partner-minjenv/kennis- finances tanken-in-vier-pitstops Hansken Politie Safety/security Investigation Implemented https://zoek.officielebekendmakingen.nl/ (criminal blg-849198.pdf activity) De Raffinaderij Politie Safety/security Investigation Unknown https://zoek.officielebekendmakingen.nl/ (criminal blg-849198.pdf activity) MyStatLab Universities Education Personalization Unknown www.overheid.nl Schengen Informatie Politie Safety/security Investigation Implemented www.overheid.nl Systeem II (SIS II) (criminal activity) (continued) The use of public sector data analytics Table A1. Type of Application (name) Organizations involved Domain application Status Source URL Image recognition: IND, TNO Migration Resource Experiment https://socialmediadna.nl/innovatiecongres- 50.000 birth allocation justitie-en-veiligheid/ certificates Logeerfraude Municipalities Safety/security Inspection and Unknown www.wrr.nl/publicaties/working-papers/ (criminal enforcement 2016/04/28/big-data-voor-fraudebestrijding activity) Diagnose Plan en Municipality of Boxmeer Safety/security Inspection and Implemented www.binnenlandsbestuur.nl/sociaal/ Sturingsinstrument (criminal enforcement achtergrond/achtergrond/het- activity) bijstandswonder-van-boxmeer.699608. lynkx Quin-technologie Pandora Intelligence and TNO Safety/security Investigation Unknown https://beveiligingnieuws.nl/nieuws/ (criminal kunstmatige-intelligentie-voorspelt- activity) vluchtroute-criminelen Alert Municipalities of Utrecht, Tilburg, Social services Inspection and Unknown Uitspraak Gerechtsh of Den Haag 28 Nijmegen, Middelburg, Hilversum, enforcement January 2014; ECLI:NL:GHDHA:2014:71 Amersfoort, Leiden, Almere, Groningen, Zwolle, Tiel, Drechtsteden SmartBox UWV, SVB, Belastingdienst, DUO, Safety/security Investigation Experiment https://zoek.officielebekendmakingen.nl/ Justid, CJIB, RDW (criminal blg-814167.pdf activity) LAA - Adreskwaliteit Ministry of BZK Data Inspection and Unknown www.rijksoverheid.nl/binaries/ BRP management enforcement rijksoverheid/documenten/rapporten/2016/ 03/21/rapport-landelijke-aanpak- adreskwaliteit/rapport-landelijke-aanpak- adreskwaliteit.pdf Monitor vitale Municipalities of Putten. Heerde, Physical Maintenance Experiment http://vitalevakantieparken.nl/default.aspx vakantieparken Epe, Harderwijk, Ede, Ermelo, environment Province of Gelderland (public sphere) (continued) TG 15,4 Table A1. Type of Application (name) Organizations involved Domain application Status Source URL AI bij meldingen Municipality of Dordrecht Physical Maintenance Experiment www.smartcitydordrecht.nl/gemeentelijke- openbare ruimte environment projecten/dordrecht-koploper-in-onderzoek- (public sphere) gebruik-artificial-intelligence-ai-bij- meldingen-openbare-ruimte Transparantie VNG Labor market Forecasting Experiment https://depilotstarter.vng.nl/projecten/ informatie over de datagedreven-werken/transparante- arbeidsmarkt informatie-over-de-arbeidsmarkt Themacontrole mbv VNG Social services Inspection and Experiment https://depilotstarter.vng.nl/projecten/ machine learning in enforcement datagedreven-werken/themacontrole-mbv- het sociaal domein machine-learning-het-sociaal-domein Opsporing Municipality of Zaanstad, VNG Safety/security Investigation Experiment https://depilotstarter.vng.nl/projecten/ ondermijning (criminal datagedreven-werken/opsporen- activity) ondermijning-met-data-analyse Brid.ge Municipality of Enschede, Saxion, Physical Personalization Experiment http://techforfuture.nl/onderzoek/afgerond/ Winkelhart Enschede, Twente environment brid-ge/ University, Both Social, eLabbs, (public sphere) Marotura, Present Media, Squal Media, Think Public Thermometer Municipality of Noordoostpolder Physical Maintenance Unknown www.destentor.nl/noordoostpolder/ vandalisme environment thermometer-voor-vandalisme-in- (public sphere) drontena375445d/ Gegevens-vergelijker Municipality of Súdwest-Fryslân Physical Inspection and Experiment https://github.com/EduardWitteveen/ environment enforcement GegevensVergelijker þ www. geoinformatienederland.nl/agenda/gin- bijeenkomst-themagroep-noord-bij-de- gemeente-sudwest-fryslan Proeftuin binnenstad- Municipality of Groningen Physical Forecasting Experiment https://groningen.nieuws.nl/gemeente/ monitoring environment 736506/innovatieve-sensoren-moeten- (mobility) binnenstad-toegankelijker-maken/ (continued) The use of public sector data analytics Table A1. Type of Application (name) Organizations involved Domain application Status Source URL Cold case Politie Safety/security Investigation Experiment www.politie.nl/nieuws/2018/mei/23/00- (criminal nieuwe-technologie-in-oude-politiezaken. activity) html Warme frontoffice Belastingdienst, DUO, CJIB, Raad Social services Personalization Experiment www.slimmernetwerk.nl/2014/05/ voor Rechtsbijstand, bureau WNSP doorbraak-doetank-schuldhulpverlening-in- zicht/ Dashboard Veiligheid Municipality of Tilburg Safety/security Investigation Experiment https://hetccv.nl/onderwerpen/ Tilburg – SpotOn informatiepositie-gemeenten/ voorbeeldgemeenten/gemeente-tilburg/ Verwijsindex VNG, Municipality of Tilburg Social services Personalization Experiment https://hetccv.nl/onderwerpen/ Risicojongeren (youth care) informatiepositie-gemeenten/ voorbeeldgemeenten/gemeente-tilburg/ þ https://vng.nl/files/vng/ 20160223_verwijsindex_risicojongeren_hv. pdf þ https://vng.nl/onderwerpenindex/ jeugd/jeugdhulp/verwijsindex- risicojongeren-vir%20 Big data Pilot Municipality of Apeldoorn Social services Investigation Experiment http://bigdatagemeenten.nl/wp-content/ Apeldoornse Jeugd (youth care) uploads/2017/10/Presentatie-Big-Data- gemeenten_27092017.pdf Scorecard data Municipalities of Enschede, Almelo, Data Maintenance Experiment https://depilotstarter.vng.nl/sites/default/ management Bloemendaal, Heemstede, Loon op management files/project_bestand/ Zand, Sudwest-Fryslan, Borger- overzicht_resultaten_publicatie_0.pptx Odoorn, Leerdam, Terneuzen, Leiden, Venray, Eindhoven, Noordoostpolder, Zoetermeer Geo-data in sociaal Municipality of Uden Social services Personalization Implemented www.digitale-sociale-kaart.nl/uden domein (continued) TG 15,4 Table A1. Type of Application (name) Organizations involved Domain application Status Source URL Website en surfgedrag Municipalities of ’s-Hertogenbosch, Not domain- Personalization Experiment https://depilotstarter.vng.nl/sites/default/ Vught, Utrecht, Veenendaal, specific files/project_bestand/rapport-exploratief- Zeewolde, Veldhoven, Den Haag, onderzoek- Heerlen, Oss, Schijndel digitaledienstverlening20met20huisstijl.pdf Viralcom – Egg Municipality of Dordrecht Physical Resource Experiment www.smartcitydordrecht.nl/projecten- environment allocation innovatiefonds/viralcom-egg (safety on the water) c6Volg Municipalities of Zwijndrecht, Social services Resource Experiment https://depilotstarter.vng.nl/sturen-op- Dordrecht (youth care) allocation resultaten/nulmeting-effectmeting-kosten- sociaal-domein þ https://depilotstarter.vng. nl/sites/default/files/project_bestand/ stapeling_in_het_ sociale_domein_de_100_van_dordrecht.pdf Geniq GGZ instellingen Social services Resource Unknown www.pinkroccade-healthcare.nl/product/ (mental health allocation geniq-ggz/ care) Dashboard Municipality of Zaanstad Social services Inspection and Experiment https://depilotstarter.vng.nl/sites/default/ slachtoffers van (mental health enforcement files/project_bestand/ huiselijk geweld care) monitor_huiselijk_geweld.docx Pronto-app Municipality of Rotterdam/ Physical Personalization Experiment https://ferm-rotterdam.nl/nl/nieuws/ Havenbedrijf Rotterdam environment havenbedrijf-rotterdam-lanceert-pronto- (infrastructure) app-voor-serieuze-efficientieverbetering Artificial inspector Rijkswaterstaat Physical Maintenance Unknown www.fabriek.ai/casus-rijkswaterstaat/ environment (infrastructure) Slimme laadpunten Provinces of Overijssel, Gelderland Physical Resource Experiment www.rtlnieuws.nl/economie/artikel/ environment allocation 4278011/gelderland-en-overijssel-plaatsen- (infrastructure) 2250-slimme-laadpalen (continued) The use of public sector data analytics Table A1. Type of Application (name) Organizations involved Domain application Status Source URL Pilot Jeugdbeleid Municipality of Rotterdam, TNO, Social services Personalization Experiment www.tno.nl/nl/aandachtsgebieden/ Rotterdam Ministry of BZK (youth care) strategische-analyses-beleid/ expertisegroepen/strategy-policy/het-policy- lab-datagedreven-beleidsontwikkeling/ Slimme Camera’s Rijkswaterstaat, Innovatiecentrale, Physical Resource Experiment https://staticresources.rijkswaterstaat.nl/ ViNotion environment allocation binaries/Project%20Slimme%20Camera% (infrastructure) 27s_tcm21-107650.pdf Digitale identificatie VNG Data Resource Experiment www.digitaleoverheid.nl/actielijn/ en verificatie management allocation onderzoek-naar-gebruik-automatische- gezichtsherkenning-en-andere-vormen-van- biometrische-identificatie/ Opsporen ’onbekende’ Municipality of Eindhoven Physical Investigation Experiment https://depilotstarter.vng.nl/projecten/ bedrijven met environment omgevingsdomein/opsporen-onbekende- milieurisico (environmental bedrijven-met-milieurisico-door-data- services) analyse Brandweerradar Brandweer, TNO Safety/security Resource Unknown www.brandweer.nl/media/4153/bom0716- allocation def-lr.pdf Chatbot Wout Politie Oost-Nederland Safety/security Personalization Experiment www.tubantia.nl/regio/politie-oost- nederland-krijgt-hulp-van-chat-robot- wouta3de02a5/ Detectiesysteem Six municipalities Safety/security Inspection and Unknown www.amersfoort.nl/nieuws/sensoren- vuurwerkoverlast enforcement helpen-vuurwerkoverlast-tegen-te-gaan.htm Amsterdam Discovery Municipality of Amsterdam, app- Physical Personalization Implemented www.computable.nl/artikel/nieuws/ Challenge-app developer environment overheid/6161734/250449/app-moet- (tourism) toeristen-verspreiden-over-amsterdam.html Rijenradar City marketing organization Physical Personalization Experiment www.computable.nl/artikel/nieuws/digital- Amsterdam Marketing, environment innovation/6002561/250449/rijenradar- Municipality of Amsterdam (tourism) toont-wachtrijen-amsterdamse-musea.html (continued) TG 15,4 Table A1. Type of Application (name) Organizations involved Domain application Status Source URL Smart Roof 2.0 Municipality of Amsterdam, Physical Maintenance Experiment https://amror.nl/smart-roof-2-0/ Waternet, Drain Products, Aedes environment Real Estate, KWR Water Cycle (water Research Institute, Marineterrein management) Amsterdam Kunstmatige Rechtbank Oost-Brabant Judiciary Resource Experiment www.rijksoverheid.nl/binaries/ Intelligentie allocation rijksoverheid/documenten/kamerstukken/ Kennissysteem voor 2018/12/19/ek-artificiele-intelligentie-en- de rechtspraak algoritmen-in-de-rechtspleging/ek- artificiele-intelligentie-en-algoritmen-in-de- rechtspleging.pdf Watson RVO RVO, Jibes Physical Personalization Experiment Eindverslag Pilot Watson RVO environment Digital we Manifestgroep, Municipality of Not domain- Personalization Experiment Presentatie bestuurlijke bijeenkomst 19 nov, Amsterdam, SIVI, Bvolve, specific 2018 Innovalor Waterbeheer SR Tauw, HAL24k Physical Maintenance Experiment Data challange inzending Tauw_Hal24k environment (water management) Statistiek van CBS, Kadaster Physical Forecasting Experiment www.kadaster.nl/-/cbs-en-kadaster-maken- nieuwbouw- environment statistiek-van-nieuwbouwtransacties? transacties redirect=%2Fweb%2Fkadaster.nl% 2Fterzake Slimme sensor Municipality of Rotterdam Physical Resource Implemented www.ad.nl/rotterdam/afvalcontainers- afvalcontainers environment allocation voorzien-van-slimme-sensoraf64633b/ Opsporing ABNAmro, University of Social services Inspection and Implemented https://nos.nl/artikel/2273390- arbeidsuitbuiting Amsterdam, Inspectie SZW enforcement mensenhandel-opgespoord-via- (mensenhandel) bankgegevens.html; www.bnr.nl/podcast/ internet-vandaag/10370844/abn-amro-jaagt- met-algoritmes-op-mensenhandel About the authors The use of Dr Anne Fleur van Veenstra is a Senior Scientist in ICT Policy at TNO’s Strategy and Policy public sector Department. Her research focuses on public sector data analytics, data-driven policy-making and data analytics public sector innovation. She has published on data-driven policy, digital government, open data and public value management. Before joining TNO, she worked as a Researcher in the field of electronic government at the Department of Technology, Policy and Management of Delft University of Technology, where she wrote her PhD thesis, entitled “IT-induced public sector transformation.” Dr Francisca Grommé is an Assistant Professor at the Organisational Dynamics of the Digital Society. Her research focuses on digitalization in relation to organizational knowledge practices: how digitalization changes the way data are collected, analyzed and verified by the state and in the digital economy. She has published about topics such as surveillance, identity, nationhood, digital work and citizenship. She previously worked at the Netherlands Organisation for Applied Scientific Research (TNO) and Goldsmiths, University of London. Her PhD thesis was written at the University of Amsterdam (AISSR) and concerns the role of technology experiments in policing. Francisca Grommé is the corresponding author and can be contacted at: gromme@essb.eur.nl Somayeh Djafari, LLM is a Senior Lawyer at the Ministry of Health, Welfare and Sport. She is also a PhD Candidate at the Law Faculty of the University of Leiden, where she is working on balancing the interests of the community as a whole and the right of the individuals affected by the legislation to respect for their private life and home in embedding the algorithm(s) by administrative authorities in the Netherlands. This article is written in a personal capacity. For instructions on how to order reprints of this article, please visit our website: www.emeraldgrouppublishing.com/licensing/reprints.htm Or contact us for further details: permissions@emeraldinsight.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Transforming Government: People, Process and Policy Emerald Publishing

The use of public sector data analytics in the Netherlands

Loading next page...
 
/lp/emerald-publishing/the-use-of-public-sector-data-analytics-in-the-netherlands-iHeLfPvwnG
Publisher
Emerald Publishing
Copyright
© Anne Fleur van Veenstra, Francisca Grommé and Somayeh Djafari.
ISSN
1750-6166
eISSN
1750-6166
DOI
10.1108/tg-09-2019-0095
Publisher site
See Article on Publisher Site

Abstract

Purpose – Public sector data analytics concerns the process of retrieving data, data analysis, publication of the results as well as re-using the data by government organizations to improve their operations and enhance public policy. This paper aims to explore the use of public sector data analytics in the Netherlands and the opportunities and challenges of this use. Design/methodology/approach – This paper finds 74 applications of public sector data analytics, identified by a Web search and consultation with policymakers. The applications are categorized by application type, organization(s) involved and application domain, and illustrative examples are used to elaborate opportunities and challenges. Findings – Public sector data analytics is most frequently used for inspection and enforcement of social services and for criminal investigation. Even though its usage is often experimental, it raises concerns for scope creep, repeated targeting of the same (group of) individuals, personal data use by third parties and the transparency of governmental processes. Research limitations/implications – Drawing on desk research, it was not always possible to identify which type of data or which technology was used in the applications that were found. Furthermore, the case studies are illustrative rather than providing an in-depth overview of opportunities and challenges of the use of data analytics in government. Originality/value – Most studies either perform a literature overview or present a single case study; this paper presents a more comprehensive overview of how a public sector uses data analytics. Keywords Data analytics, Big data, Open data, Artificial intelligence, Evidence-based policy-making, Data-driven policy-making Paper type Research paper © Anne Fleur van Veenstra, Francisca Grommé and Somayeh Djafari. Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) licence. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and non-commercial purposes), subject to full attribution to the original publication Transforming Government: People, Process and Policy and authors. The full terms of this licence may be seen at http://creativecommons.org/licences/by/4.0/ legalcode pp. 396-419 Emerald Publishing Limited This paper is based on the study “Quickscan AI in de publieke dienstverlening,” which was 1750-6166 DOI 10.1108/TG-09-2019-0095 commissioned by the Ministry of the Interior and Kingdom Relations of the Netherlands. 1. Introduction The use of Following developments in the private sector, governments around the world use data public sector analytics to capture the value of data and improve their way of working (Kim et al.,2014). data analytics Public sector data analytics concerns the data retrieval process, analysis of data, publishing the results as well as re-using the data to address societal challenges (Klievink et al.,2017; Lnenicka and Komarkova, 2019; Ooijen et al.,2019). It is driven by the increased availability of open and big data and data processing techniques (Janssen and Kuk, 2016; Van Veenstra and Kotterink, 2017). Furthermore, attention to data-driven methodologies is growing as governments aim to improve the effectiveness of their policies by making them more evidence-based (Höchtl et al.,2016; Giest, 2017; Poel et al., 2018). The use of public sector data analytics, however, has also led to concerns for the privacy of individuals and the transparency of government operations and decisions (Janssen and van den Hoven, 2015). Studies of public sector data analytics in the literature on digital government often focus on data availability for public or economic value (Susha et al.,2015; Millard, 2018). Actual usage has, so far, attracted much less attention and empirical evidence on which value is created or which concerns prove to be grounded is fragmented (Jagadish, 2015; Janssen and Kuk, 2016). Not surprisingly, few comprehensive empirical studies of data analytics in the public sector have been carried out; instead, research often presents a literature overview (cf. Mehr, 2017; Wirtz et al., 2019; Lnenicka and Komarkova, 2019) or individual case studies (cf. Zheng et al., 2018; Androutsopoulou et al., 2019; Sun and Medaglia, 2019). Therefore, we aim to contribute to closing this gap in the literature by investigating usage in the Netherlands based on a large number of applications. By analyzing current use of public sector data analytics including opportunities and challenges, we substantiate the relevance of these opportunities and challenges. A study of the Netherlands, a digital frontrunner within Europe (cf. its third ranking in the Digital Economy and Society Index [1]), is expected to yield state-of-the-art insights. Although the public sector has used data analytics for decades, we aim to investigate current developments and, therefore, focus on current usage and usage that ended in 2015 or later (also see the method section regarding selection and identification). Based on a Web search, we identify and analyze 74 applications in the public sector (see Appendix). We then further elaborate on the main findings using illustrative examples from Dutch practice. Before presenting these data, the next section provides a review of previous literature on the topic. We end the paper with a discussion, followed by conclusions and recommendations for further research. 2. Public sector data analytics To investigate the use of public sector data analytics, we first look at its drivers, followed by an overview of different types of use, and opportunities and challenges. The use of public sector data analytics is driven on the one hand by technological developments, such as the increased availability of data for re-use and the advance of data processing technologies. Data availability is increased by the publication and re-use of governmental data as open data (Janssen et al., 2012) and by the combination of a variety of data sets from different organizations (including the private sector) as big data (McAfee and Brynjolfsson, 2012; Manyika et al., 2011). Data processing technologies includes artificial intelligence (AI), which refers to systems that solve problems, take decisions or carry out tasks that require intelligence (Russell and Norvig, 2010) enable government organizations to capture the value of the available data. On the other hand, public sector data analytics is driven by government organizations aiming to address societal issues and inform the different steps of policy-making by making it more evidence-based (Höchtl et al.,2016; Giest, 2017). Following these two drivers also makes it possible to discern different types of use. TG Several studies identified (technical) applications of data analytics (Mehr, 2017; Chui et al., 15,4 2018; Wirtz et al., 2019). Mehr (2017), for example, categorizes the types of (technical) government problems appropriate for the use of AI, which include resource allocation, large data sets, a shortage of expertise, predictable scenarios and procedural and diverse data. Based on an analysis of advanced usage in different industries, Chui et al. (2018) find three main categories to improve performance: predictive maintenance, logistics optimization and personalization. And a list of ten potential technical applications of analytics in the public sector is provided by Wirtz et al. (2019), ranging from process automation to speech analytics. These types of usage focus on technical possibilities, but do not link to specific roles of government organizations. Others have identified types of usage that aim to improve policy-making (Santiso and Roseth, 2018; Poel et al., 2018; Ooijen et al., 2019). Santiso and Roseth (2018) and Ooijen et al. (2019) distinguish four types of analytics – descriptive, diagnostic, predictive and prescriptive – for different steps of policy-making. Poel et al. (2018) also distinguish different phases of policy-making, including foresight and agenda setting, monitoring and interim evaluation, problem analysis, identification and design of policy options, policy implementation and ex post evaluation and impact assessment. Furthermore, other types of usage are found by looking at individual illustrations of its use in public service delivery (Zheng et al.,2018; Androutsopoulou et al.,2019). Few of these studies have focused on actual usage, and none have looked at which types of governmental organizations use data analytics, in which domain, and which societal challenges are addressed [with the exception of Poel et al. (2018) that listed different policy areas]. Opportunities that are associated with the use of public sector data analytics are making governments’ operations more efficient and effective (Janssen and Kuk, 2016; Klievink et al., 2017) and increasing transparency, which is linked to openness and accountability of governmental processes (Meijer, 2009; Matheus and Janssen, 2015). Furthermore, opportunities for the use of data analytics for policy making are anticipatory governance, design and delivery and performance management (Ooijen et al.,2019). There are, however, also challenges associated with the use of public sector data analytics. Technical challenges are ensuring data quality and security (Bertot and Choi, 2013; Janssen et al.,2012). Also the re-use of data requires the presence of a data infrastructure ensuring interoperability (Lnenicka and Komarkova, 2019; Ooijen et al.,2019), and attention should be given to linking data semantically (Bizer, 2009), to allow for findability and re-use. Besides linking data technically, many studies have found that merely publishing data is insufficient to enable and stimulate re-use and, therefore, call for supporting collaborations with relevant stakeholders (Susha et al.,2017; Styrin et al.,2017) as well as for co-creation with citizens (Ferro et al.,2013; Charalabidis et al.,2014; van Veenstra and Kotterink, 2017). Challenges associated with the use of data analytics for policy-making are the public sector’s limited organizational readiness (Klievink et al., 2017), including governments’ hesitance to rely on new or emerging data sources (Poel et al.,2018), decision-makers’ unwillingness to be informed by data science (van der Voort et al.,2019) and negative consequences for operational employees (Barocas and Selbst, 2016). Furthermore, research shows that the expectation of many government organizations to become transparent as a result of using data analytics too is often high (Bannister and Connolly, 2012). The emergence of bias as a result of the use of algorithms is also considered a major challenge (Janssen and Kuk, 2016). Bias among others originates in the training data used for developing and refining algorithms, and can eventually affect decision-making (Hacker, 2018). Finally, legal concerns include privacy and accountability of governmental decision- making, for instance, when the application of a technology changes scope (“scope creep”) The use of and ceases to be proportional (Janssen and van den Hoven, 2015; Lourenço et al.,2015; public sector Zoonen, 2016, 2019; Agarwal, 2018). data analytics This literature review makes clear that public sector data analytics is driven by the application of innovations that have the potential to improve policy-making and address societal challenges, but many technical as well as organizational and legal challenges exist and it is not yet clear how these opportunities and challenges are related to actual usage. 3. Method The investigation of public sector data analytics in the Netherlands took place between November 2018 and March 2019. For the identification of applications of public sector data analytics, a structured Web search resulted in 74 examples, in which we included different government levels (municipal, provincial and national) as well as public and semi-public agencies. The Web search covered (academic) search engines and key government databases [2]; these were searched using the terms “AI,”“AI AND government,”“Data AND government” and “Algorithm AND government.” Google, Google Scholar and PiCarta (the national library catalogue) were used for orientation and complementary information. Another source for orientation was LexisNexis, a research database covering a large range of print media in the Netherlands. For up-to-date information on public usage, we relied on the following government and semi-government databases: Overheid.nl (the main portal for government information) and the Pilot Starter database, which lists pilot projects conducted by municipalities. Overheid.nl is a central access point to all government information. We therefore searched this database for additional terms in full-text: “kunstmatige intelligentie” (AI), “experiment” and “pilot.” Using this combination of sources, several characteristics were captured and verified for each application, including application type, organization type, status (implemented or experimental) and domain of usage. These characteristics were chosen because they capture the nature and aims of usage (application type) or because they have not yet been structurally investigated (type of organization and domain of usage). The categorization of applications was performed by two researchers independently; when differences between the categorizations occurred, a discussion followed based on which a decision was made. The Web search results were reviewed by a panel of policymakers from different levels of government. This panel was formed by two representatives from the Ministry of the Interior, two from the Association of Dutch Municipalities and six from different municipalities (all involved in using data analytics). The review took place as a session during a workshop and was facilitated by one author of this paper, while another author took notes. The purpose of the session was to validate whether the applications represent types of use in the Netherlands. This panel also provided information on the current status of some applications and provided internal policy documents that helped to specify the types of data and analytics that were used (the latter details are often difficult to verify through a Web search). Finally, we acknowledge that the resulting overview cannot provide a comprehensive record of all instances of public sector data analytics. First, we needed to balance the identification of applications with collecting and verifying the characteristics of each project. Furthermore, not all applications are made public online and we expect our search terms to miss a number of initiatives. In addition, public sector data analytics is a dynamic field which limits the extent to which any list can capture all usage. Not knowing which applications we have missed, it is not possible to determine how representative this list is. Nevertheless, the chosen government databases and additional sources do provide a basis for identifying key types of usage and characteristics. We use illustrative examples (based TG on information from public documents, such as policy documents and newspaper articles) to 15,4 elaborate on the main findings on these types of usage and characteristics (see Section 5). 4. Analysis of the applications Based on the list of 74 applications, three analyses were undertaken: a categorization of the examples by application type and status; by type of organization and application type; and by application domain and type. 4.1 Types of applications To the best of our knowledge, there is no established categorization of public sector data analytics that captures types of use. Our literature review found technological and policy- making oriented categorizations, but none focusing on of usage. Therefore, we categorized the 74 applications into 6 types of usage, inspired by the literature overview and refined based on the applications found online: (1) Personalization: the tailoring of public service delivery to individual needs. (2) Resource allocation: optimizing logistics for efficient use of resources. (3) Maintenance: identifying potential risks to schedule repairs. (4) Inspection and enforcement: predicting (behavioral) patterns to prevent misuse of public spending. (5) Crime investigation: pattern recognition and prediction such as fraud. (6) Forecasting: predicting macro-economic trends and variables. These types of usage are a mix of operational issues that government organizations may wish to address in different phases of the policy-making process (problem definition, execution and monitoring) as well as for other roles government organizations (operations and service delivery) have. We found experimental applications, meaning that they are not yet part of the standard operational activities of government, as well as implemented practices that are fully operational and part of the standard activities of government organizations. Regarding the categorization shown in Figure 1, the category with the largest number of data analytics applications in our list of examples is “inspection and enforcement” (20 out of Figure 1. The application domains per type of application 74). This category also includes the largest group of implemented applications (8 out of 20). The use of A total of 17 out of 74 are categorized as “investigation”; 4 out of these 17 were found to be public sector implemented. A total of 11 out of 74 applications fit the category of “resource allocation” and data analytics 9 applications are categorized as “maintenance.” For the latter two categories, only one implemented application was found (for “resource allocation”). Of the 74 applications, 14 are categorized as “personalization”; 2 out of these 14 were implemented. Finally, the category for which we found the lowest number of applications is “forecasting” (3 out of 74); all 3 are experiments. Although the number of applications for each category can be an effect of the search strategy and availability, the findings are indicative of the important role of experimentation in all categories. It also suggests that inspection and enforcement, investigation and maintenance and resource allocation are active fields for experimentation and implementation. 4.2 Types of government organizations The types of government organizations (municipality, central government, province or water board [regional], collaboration, executive or semi-public organization or other) involved in the use of data analytics are presented in Figure 2. Of the 74 applications, 18 we found are collaborations between different levels of government or between public and semi-public or non-governmental organizations, such as companies and non-governmental organizations. Another 16 applications are collaborations between municipalities or regional governments and are categorized as municipal or regional. Apart from forecasting, we found a few (2–5 out of 18) applications for every application type within the identified collaborations, indicating collaborations are formed for various types of applications. We found municipalities and executive and semi-public organizations to each be involved in 24 of the 74 applications, while 5 applications were found for “other,” 2 applications for the national and 1 for the regional level. Although national and regional governments are also represented in several collaborations and the exact proportions may not reflect all current activity, these findings do suggest the active role of municipalities in public sector data analytics in the Netherlands. The applications we found on the municipal level most often concern “inspection and enforcement” (8 out of 25), “personalization” (5 out of 25) and “investigation” and “resource allocation” (both 4 out of 25), while we found fewer applications for “maintenance” and Figure 2. Types of government organizations per type of application “forecasting.” Within the group of applications of executive and semi-public organizations, 9 TG out of the 24 applications are categorized as “investigation” and 5 out of the 24 as 15,4 “inspection and enforcement.” Only two or three applications were found for the other categories. The latter cover all types except for forecasting, for which no applications were found. 4.3 Application domain To investigate which sectors use data analytics, finally, we focus on their application domain (Figure 3). Most applications were found in the physical environment, which includes the public sphere, water management, mobility and infrastructure (22 out of 74). In this domain, all types of applications of data analytics were found. Most applications were found in the categories “maintenance” (7 out of 22), “personalization” (5 out of 22) and “resource allocation” (5 out of 22). In the physical domain, usage can thus be considered to comprise all application types. Two other domains in which many applications were found are safety and security (20 out of 74) and social services (17 out of 74). Within the safety and security domain, the applications we found are overwhelmingly categorized as “investigation” (15 out of 20); in the social domain, a majority of applications concerned “inspection and enforcement” (10 out of 17). Compared to the physical domain, data analytics usage in these domains is more focused. For the other domains (data infrastructure and finance, which includes applications linked to the governmental data infrastructure and finance, economy, labor market and education; and other, comprising applications from the legal domain and in education), few applications were found and with various application types. Therefore, no uniform conclusions can be drawn for these domains. 5. Key findings, opportunities and challenges of data analytics In this section, we discuss four main findings based on the overview in Section 4. Our aim is to relate these findings to opportunities and challenges of public sector data analytics. The opportunities and challenges discussed in the context of each finding serve to indicate that research focused on gaining an overview of the field can also contribute to insights into how opportunities and challenges of data analytics emerge. These brief discussions are not Figure 3. The types of application and their implementation meant to be comprehensive, instead, we highlight one or two opportunities and challenges in The use of relation to each finding on the basis of prominent themes in the literature (Section 2). As public sector noted in the literature review, among the opportunities are increasing efficiency, data analytics effectiveness and transparency. Transparency also is a relevant challenge related to data analytics, as are legal concerns (privacy and accountability), bias and organizational readiness. A disadvantage of an overview study is that it does not support a more detailed understanding of types of projects that have been executed and how they develop in practice. This section therefore also presents illustrative examples of applications from the list of 74, drawing on publicly available documents. Without claiming to be generalizable, these examples (sensors in municipal waste containers; data mining for social welfare fraud; a chatbot; and data mining for crime investigation) were chosen because they represent typical uses of data analytics in the Dutch public sector, and we therefore expect them to be informative for our understanding of future developments. 5.1 Many applications are used for operational purposes Many applications found are for operational usage, such as resource allocation and inspection and enforcement. Forecasting was the least common usage we found. The opportunity of using data analytics for everyday operations of public services is exemplified by the placement of sensors in household waste containers by the municipality of Rotterdam to improve resource allocation and maintenance (Gemeente Rotterdam, 2020). A total of 6,500 sensors were used to compute the most efficient collection route, thereby not only attempting to increase the quality of services but also to decrease CO emissions. Furthermore, the sensors also keep track of container defects. After a pilot in 2017, the system has been scaled to the city level in 2018 (Gemeente Rotterdam, 2017). A challenge characteristic for this type of usage is the collection of data about mundane aspects of citizens’ lives by third parties that carry out activities for public service delivery. This is exemplified by the Rotterdam waste service, which relies on third parties for data collection and processing. For other operational services, some critics have noted that third party involvement is not only a privacy challenge but can also affect the transparency and accountability of public services as it can be difficult for citizens to find and address the third parties involved (Van Teeffelen and Naafs, 2017). 5.2 Most often identified applications concern “inspection and enforcement” Inspection and enforcement is the most common usage category among the identified applications. Furthermore, half of these applications were found in the social domain. The use of data analytics is, thus, often seen as an opportunity to efficiently prevent misuse of government funds and decrease social service fraud. Systeem Risico Indicatie (SyRI) or “risk indication system” is an example of the promise of data analytics to this end. Discontinued in 2020, SyRI was a data mining application combining a variety of data types, such as data on taxes, properties, education, pension, debt, benefits and health insurance to search for anomalies and create risk profiles (Olsthoorn, 2016). The system was used from 2014 by varying constellations of government organizations, including municipalities, the Inland Revenue and the Inspection of the Ministry for Social Affairs. An important concern relating to inspection and enforcement is its reliance on the analysis of large amounts of sensitive data. A risk is that such data can be collected and analyzed for large groups of people without specific grounds for suspicion. This risk is related to a starting point of many data mining methods: a more open-ended analysis of TG large amounts of data to find previously unknown behavioral patterns (Olsthoorn, 2016). In 15,4 the case of SyRI, civil rights advocates started a lawsuit against the state because of such practices, which they referred to as “fishing expeditions” (Platform Bescherming Burgerrechten, 2018). In February 2020, court agreed that such data usage is not proportional and ruled that SyRI was in breach of Article 8 of the EU Charter of Human Rights (the right to privacy). On top of this, court ruled that the SyRI program lacked transparency and accountability, as it did not provide access to the algorithm that generated risk profiles. A second concern relevant to highlight regarding inspection and enforcement is that data analytics practices may display a bias toward vulnerable groups. This also characterizes the social domain, as governments often collect more data about these groups (Barocas and Selbst, 2016; Janssen and Kuk, 2016). SyRI, for instance, predominantly targeted people in low-income neighborhoods (about whom more data are available). 5.3 Many applications concern experimental usage Over half of the applications found are pilots or experiments (even though for many applications, the implementation status could not be confirmed). Experimentation provides degrees of freedom needed to apply and develop a technology and explore its capacities. For instance, a chatbot piloted by the national service center for entrepreneurs (Rijksdienst voor Ondernemend Nederland, [RVO]) was used as a search engine for its website using natural language processing. The aim of this 2017 experiment was to facilitate entrepreneur compliance with complex legislation regarding agricultural manure policies (relevant to nitrogen emissions). The chatbot prototype was further developed and integrated into Facebook Messenger to redirect questions away from the call center (RVO, 2017). While personalization offers opportunities to improve service delivery toward citizens, to tailor (often complex) government information to users and to cut costs on service delivery, a key challenge is to ensure accessibility because it involves direct interaction between information technology systems and citizens. The RVO chatbot is a good illustration of the problems that can emerge when this is not successful. During the development of the first version of the system, the RVO testers commented that the chatbot came across as “robotic,” thereby lacking human nuance (RVO, 2018). Furthermore, call center employees were not able to ensure accessible service delivery at first because they could not assume control of the system. Another opportunity of experimentation is to enroll stakeholders without demanding long-term commitment. FinPro is an example of criminal investigation that took advantage of the experimental format’s flexibility. The name refers to a collaboration established for data experiments by platform organization MeetingMoreMinds in 2011, which included, among others, the Rotterdam police, the Chambers of Commerce, health insurance companies, travel organizations and bus companies (the collaboration has not formally been disbanded, although most activities were conducted before 2017). The main purpose was to identify behavioral patterns on an aggregated level. The experimental format allowed for trialing the ad hoc involvement of experts from different domains to analyze data and interpret findings; a feature that initially led to new insights (Roobeek, 2016). However, the use of experimental data analytics may entail transparency issues as not all pilots are well documented and made available to public scrutiny. 5.4 Many data analytics projects are executed by multi-stakeholder collaborations The use of Nearly half of the identified applications concern collaborations: between the same type of public sector organizations, between different levels of government or between public and semi-public or data analytics non-governmental organizations. Collaborations allow for combining data sets to identify new patterns or anomalies in the data, and for attracting financial and technical resources, as illustrated by FinPro (Roobeek, 2016). Yet, a challenge of collaboration is “scope creep”: when the application of a technology broadens its scope beyond its original intensions. This also means that legal concerns arise that were not identified in advance. For instance, the FinPro collaborators learned along the way that the outcomes of data analysis can often not be submitted as legal evidence (Jessayan and De Lange, 2016). To return to the example of SyRI discussed above, enlarging the number of involved stakeholders entailed multiplying the number of interests and aims invested in the project. In SyRI, the scope was broadened over the years from welfare fraud to include other social issues such as harassment and nuisance (Raad van State, 2014). Scope creep thus raises legal concerns as the purpose of data collection and analysis may not be proportionate to the infringement of personal sphere. To conclude, we elaborated on four main findings to indicate that research focused on improving our overview of the field can also contribute to insights to how opportunities and challenges of data analytics emerge; and support more detailed insights into types of projects that have been executed and how they develop in practice. We further discuss our findings in the next section. 6. Discussion Our four main findings both corroborate and contribute to current research on public sector data analytics. First, we found that many applications aim for efficiency and effectiveness of government operations rather than for increasing transparency or facilitating evidence-based policy-making. This corroborates findings from Giest (2017) and Poel et al. (2018) that the use of data analytics for policy is still in its early days. Furthermore, third parties’ data analytics activities in operational processes of the public sector lead to concerns of third party access to data about everyday citizen behavior and of transparency and accountability of these processes. These concerns have been amply addressed in the literature about smart cities (Zoonen, 2016) but are equally relevant to research on digital government. Second, a frequent use type among the applications we identified is inspection and enforcement for social services (often used by municipalities). This implies that, within operational applications, data analytics is often used for investigation or enforcement of legislation. This makes the societal, ethical and legal concerns connected to this use type key in studies of public sector data analytics [see, for example, Janssen and van den Hoven (2015)]. Next to proportionality and invasiveness, the repeated targeting of (groups of) individuals based on their socio-economic class (or that of their neighborhood) also emerged as a concern. We know from the wider literature on the risk of bias in big data usage that when targeted more often, citizens are also at greater risk of becoming “anomalies in the data” and objects of further investigation (Barocas and Selbst, 2016), with an associated risk of stigmatization. Third, as more than half of the identified applications are experiments, this implies that, as suggested by the literature, many applications of data analytics face difficulties moving beyond the experimentation phase because of technical and legal challenges [see, for instance, van Veenstra and Kotterink (2017)]. Similarly, personalization and tailoring of services still seem to be under development, implying that data analytics are neither easily deployed effectively for this purpose nor lead to more accessibility in the experience of TG citizens. As emphasized in the literature (Klievink et al.,2017), to gain from the opportunities 15,4 that these technologies may offer, attention also needs to be given to organizational readiness and change. As well as experimental, many projects are collaborative, which is our fourth finding. From two of our illustrative examples, we furthermore learn that these collaborations can involve changing stakeholder constellations. Operating in a multi- actor context, as is the case for many data analytics initiatives in the Netherlands, is an additional challenge for guarding the scope of projects. Furthermore, this coincides with an observation regarding the applications we identified: we find that much attention is given to data sharing among different stakeholders rather than to using more advanced processing technologies such as AI (Sun and Medaglia, 2019); applications of collaborative data analytics often combine data sets and search for behavioral patterns or anomalies. While this study was one of a very few providing an overview of public sector data analytics based on a larger number of cases, a main limitation is its reliance on desk research. Therefore, it was often not possible to present details such as the type of analytics that was used. Consequently, it is impossible to draw conclusions based on this study about which specific technologies are used by government. A first avenue for further research, therefore, is to examine which technologies are used. A second avenue is examining the opportunities and challenges of public sector data analytics more structurally and in-depth, capturing also the views of those involved. 7. Conclusion This study explores the use of public sector data analytics in the Netherlands and aims to contextualize opportunities and challenges identified in previous research. Based on a Web search, we find that many identified applications are used for operational processes, and are often deployed in processes of inspection and enforcement for social services. Furthermore, many applications are pilots or experiments rather than routine usage and they concern collaborations between different stakeholders. We thus find that advance data processing is still in its early stages. At the same time, the illustrative examples raise several concerns including third-party access to personal data, repeated targeting of the same (groups of) individuals, limitations to accessibility and transparency and scope creep, especially in multi-actor settings. This implies that the use of public sector data analytics requires developing organizational capabilities to ensure effective use, foster collaboration and scale up, as well as legal and ethical capabilities to carefully balance these concerns with their objectives, such as increased operational efficiency or effectiveness. As a major limitation of this study is its reliance on desk research, further research should perform more in-depth research structurally capturing the opportunities and challenges of the use of the application types of public sector data analytics. Notes 1. The Digital Economy and Society Index is a composite index with digital public services as one of the components. For more information, see: https://ec.europa.eu/digital-single-market/en/desi 2. The source URLs are: google.com; scholar.google.com; picarta.nl; lexisnexis.nl; overheid.nl; and depilotstarter.vng.nl. References The use of Agarwal, P.K. (2018), “Public administration challenges in the world of AI and bots”, Public public sector Administration Review, Vol. 78 No. 6, pp. 917-921. data analytics Androutsopoulou, A., Karacapilidis, N., Loukis, E.N. and Charalabidis, Y. (2019), “Transforming the communication between citizens and government through AI-guided chatbots”, Government Information Quarterly, Vol. 36 No. 2, pp. 167-384. Bannister, F. and Connolly, R. (2012), “The trouble with transparency: a critical review of openness in e- Government”, Policy and Internet, Vol. 3 No. 1, pp. 1-30. Barocas, S. and Selbst, A.D. (2016), “Big data’s disparate impact”, California Law Review, No. 104, pp. 671-732. Bertot, J.C. and Choi, H. (2013), “Big data and E-government: issues, policies, and recommendations”,in Proceedings of the 14th Annual International Conference on Digital Government Research. dg.o. 2013, 17–20 June, Quebec City, Canada, pp. 1-10. Bizer, C. (2009), “The emerging web of linked data”, IEEE Intelligent Systems, Vol. 24 No. 5, pp. 87-92. Charalabidis, Y., Loukis, E.N., Androutsopoulou, A., Karkaletsis, A. and Triantafillou, A. (2014), “Passive crowdsourcing in government using social media”, Transforming Government: People, Process and Policy, Vol. 8 No. 2, pp. 283-308. Chui, M., et al. (2018), Notes from the AI Frontier: Applications and the Value of Deep Learning Insights from Hundreds of Cases, McKinsey Global Institute, available at: www.mckinsey.com//media/ McKinsey/Featured%20Insights/Artificial%20Intelligence/Notes%20from%20the%20AI%20 frontier%20Applications%20and%20value%20of%20deep%20learning/Notes-from-the-AI- frontier-Insights-from-hundreds-of-use-cases-Discussion-paper.ashx Ferro, E., Loukis, E.N., Charalabidis, Y. and Osella, M. (2013), “Policy making 2.0: from theory to practice”, Government Information Quarterly, Vol. 30 No. 4, pp. 359-368. Gemeente Rotterdam (2017), “Persbericht: Minder kosten en CO2 uitstoot door slimme afvalcontainers”, Rotterdam.nl, available at: www.persberichtenrotterdam.nl/bericht/2270/Minder-kosten-en-CO2- uitstoot-door-slimme-afvalcontainers/ Gemeente Rotterdam (2020), “Slimme wijkcontainers”, Rotterdam.nl, available at: www.rotterdam.nl/ wonen-leven/slimme-wijkcontainers/ Giest, S. (2017), “Big data for policymaking: fad or fasttrack?”, Policy Sciences, Vol. 50 No. 3, pp. 367-382. Hacker, P. (2018), “Teaching fairness to artificial intelligence: Existing and novel strategies against algorithmic discrimination under EU law”, Common Market Law Review, Vol. 55 No. 4, pp. 1134-1185. Höchtl, J., Parycek, P. and Schöllhammer, R. (2016), “Big data in the policy cycle: Policy decision making in the digital era”, Journal of Organizational Computing and Electronic Commerce, Vol. 26 Nos 1/2, pp. 147-169. Jagadish, H.V. (2015), “Big data and science: Myths and reality”, Big Data Research, Vol. 2 No. 2, pp. 49-52. Janssen, M. and Kuk, G. (2016), “The challenges and limits of big data algorithms in technocratic governance”, Government Information Quarterly, Vol. 33 No. 3, pp. 371-377. Janssen, M. and van den Hoven, J. (2015), “Big and open linked data (BOLD) in government: a challenge to transparency and privacy?”, Government Information Quarterly, Vol. 32 No. 4, pp. 363-368. Janssen, M., Charalabidis, Y. and Zuiderwijk, A. (2012), “Benefits, adoption barriers and myths of open data and open government”, Information Systems Management,Vol.29 No. 4, pp. 258-268. Jessayan, H. and De Lange, B. (2016), “Big data legt onzichtbare criminaliteit bloot”, Het Financieele Dagblad, 9 June. Kim, G.-H., Trimi, S. and Chung, J.-H. (2014), “Big data applications in the government sector”, TG Communications of the Acm, Vol. 57 No. 3, pp. 78-85. 15,4 Klievink, B.,Romijn,B.-J.,Cunningham,S., Bruijn,H.and de,(2017), “Bigdata inthe public sector: Uncertainties and readiness”, Information Systems Frontiers,Vol.19 No. 2, pp. 267-283. Lnenicka, M. and Komarkova, J. (2019), “Big and open linked data analytics ecosystem: Theoretical background and essential elements”, Government Information Quarterly, Vol. 36 No. 1, pp. 129-144. Lourenço, R.P., Piotrowski, S. and Ingrams, A. (2015), “Open data driven public accountability”, Transforming Government: People, Process and Policy, Vol. 11 No. 1, pp. 42-57. McAfee, A. and Brynjolfsson, E. (2012), “Big data: the management revolution”, Harvard Business Review, No. October, pp. 1-9. Manyika, J., Chui, M., Brown, B., Bughin, J., Dobbs, R., Roxburgh, C. and Hung Byers, A. (2011), Big Data: The Next Frontier for Innovation, Competition, and Productivity, McKinsey Global Institute, available at: www.mckinsey.com/business-functions/mckinsey-digital/our-insights/ big-data-the-next-frontier-for-innovation Matheus, R., Janssen, M., et al. (2015), “Transparency dimensions of big and open linked data transparency as being synonymous with accountability and openness”, in: Janssen (Eds.) I3e 2015, Lncs 9373, Springer, pp. 236-246. Mehr, H. (2017), “Artificial intelligence for citizen services and government”, Harvard Ash Center, available at: http://ash.harvard.edu/files/ash/files/artificial_intelligence_for_citizen_services. pdf, Meijer, A. (2009), “Understanding modern transparency”, International Review of Administrative Sciences, Vol. 75 No. 2, pp. 255-269. Millard, J. (2018), “Open governance systems: Doing more with more”, Government Information Quarterly, Vol. 35 No. 4, pp. 77-87. Olsthoorn, P. (2016), “Big data voor fraudebestrijding”, WRR Working paper no. 21, available at: www. wrr.nl/publicaties/working-papers/2016/04/28/big-data-voor-fraudebestrijding Ooijen, C., van Ubaldi, B. and Welby, B. (2019), “A data-driven public sector”, OECD Working Papers on Public Governance No. 33. Platform Bescherming Burgerrechten (2018), “NGO’s starten rechtszaak tegen de staat over risicoprofilering van nederlandse burgers”, December 1, available at: https://platform burgerrechten.nl/2018/01/12/ngos-starten-rechtszaak-tegen-de-staat-over-risicoprofilering-van- nederlandse-burgers/ Poel, M., Meyer, E.T. and Schroeder, R. (2018), “Big data for policymaking: Great expectations, but with limited progress?”, Policy and Internet, Vol. 10 No. 3, pp. 347-367. Raad van State (2014), “Advies W12.14.0102/III”, available at:www.raadvanstate.nl/adviezen/advies. html?id=11339 Roobeek, A. (2016), “Finpro geeft met verantwoorde Big-Datamethoden zicht op onzichtbare veiligheid”, Tijdschrift Voor de Politie, Vol. 78 No. 9, pp. 24-28. Russell, S.J. and Norvig, P. (2010), Artificial Intelligence: A Modern Approach, Third edition, Pearson Education, Upper Saddle River, NJ. RVO (2017), Eindrapport Pilot IBM Watson, The Hague, RVO. RVO (2018), Presentatie: Bestuurlijke Bijeenkomst Artificial Intelligence (AI)/Chatbot, The Hague, RVO. Santiso, C. and Roseth, B. (2018), “Data disrupts corruption”, Stanford Social Innovation Review, Spring, pp. 50-55. Styrin, E., Luna-Reyes, L.F. and Harrison, T.M. (2017), “Open data ecosystems: an international comparison”, Transforming Government: People, Process and Policy,Vol.11 No. 1, pp. 132-156. Sun, T.Q. and Medaglia, R. (2019), “Meeting the challenges of artificial intelligence in the public The use of sector: Evidence from public healthcare”, Government Information Quarterly, Vol. 36 No. 2, public sector pp. 368-383. data analytics Susha, I., Grönlund, A. and Janssen, M. (2015), “Organizational measures to stimulate user engagement with open data”, Transforming Government: People, Process and Policy, Vol. 9 No. 2, pp. 181-206. Susha, I., Janssen, M. and Verhulst, S. (2017), “Data collaboratives as “bazaars?”, Transforming Government: People, Process and Policy, Vol. 11 No. 1, pp. 157-172. van der Voort, H.G., Klievink, A.J., Arnaboldi, M. and Meijer, A.J. (2019), “Rationality and politics of algorithms. will the promise of big data survive the dynamics of public decision making?”, Government Information Quarterly, Vol. 36 No. 1, pp. 27-38. Van Teeffelen, K. and Naafs, S. (2017), “Moeten We wel zo blij zijn met de slimme stad?”, Trouw,5 December, available at: www.trouw.nl/samenleving/moeten-we-wel-zo-blij-zijn-met-de-slimme- stad-aaa1a488/ Van Veenstra, A.F., Kotterink, B., et al. (2017), “Data-Driven policy making: the policy lab approach”,in: Parycek (eds.) ePart 2017, LNCS 10429, Springer, pp. 100-111. Wirtz, B.W., Weyerer, J.C. and Geyer, C. (2019), “Artificial intelligence and the public sector – applications and challenges”, International Journal of Public Administration, Vol. 42 No. 7, pp. 596-615. Zheng, Y., Yu, H., Cui, L., Miao, C., Leung, C. and Yang, Q. (2018), “SmartHS: an AI platform for improving government service provision”, Association for the Advancement of Artificial Intelligence, January 22, available at: www.ntulily.org/wp-content/uploads/conference/Smart HS_An_AI_Platform_for_Improving_Government_Service_Provision_accepted.pdf Zoonen, L.V. (2016), “Privacy concerns in smart cities”, Government Information Quarterly, Vol. 33 No. 3, pp. 472-480. Zoonen, L.V. (2019), “Opnieuw fatale remedies”, Sociologie, Vol. 15 No. 1, pp. 19-43. Appendix TG 15,4 Table A1. Applications of public sector data analytics in The Netherlands Type of Application (name) Organizations involved Domain application Status Source URL SyRI UWV, SVB, Inspectie SZW, Social services Inspection and Implemented https://wetten.overheid.nl/BWBR0013060/ Belastingdienst and municipalities enforcement 2019-01-01 Criminaliteits Politie Safety/security Investigation Unknown www.politie.nl/nieuws/2017/mei/15/05-cas. Anticipatie Systeem html (CAS) Infobox crimineel en Openbaar Ministerie, Politie, Safety/security Investigation Implemented Stcrt. 2015, nr. 7509 and https://hetccv.nl/ onverklaarbaar Belastingdienst, FIOD and onderwerpen/veilige-vakantieparken/ vermogen (iCOV) Financial Intelligence Unit, overzicht-maatregelen/analyse-icov/; www. Inspectie SZW, Inlichtingen- en securitymanagement.nl/big-data-fraude/? Opsporingsdienst van de Inspectie vakmedianet-approve-cookies=1&_ga= voor de Leefomgeving en Transport 2.167311545.627848040.1548922282- (ILT-IOD), Inlichtingen- en 359266002.1548922282 Opsporingsdienst van de Nederlandse Voedsel- en Warenautoriteit (NVWA-IOD), Commissariaat voor de Media iRN_iColumbo Politie Safety/security Investigation Unknown TK 2013–2014, 26 643, nr. 298, p. 16 Business Intelligence Belastingdienst Government Inspection and Implemented https://decorrespondent.nl/1766/Vergeet-de- and Analytics (BI&A) finances enforcement politiestaat-Welkom-in-de-belastingstaat/ 90525160-5a2c27c2 Verwijzingsindex Financial Intelligence Unit Safety/security Investigation Implemented https://wetten.overheid.nl/BWBR0041193/ Recherche Nederland (FIU-Nederland) 2018-07-25 Onderzoeken en Subjecten (VROS- bestand) Parkeersensoren voor Municipality of Zoetermeer Physical Resource Experiment https://depilotstarter.vng.nl/projecten/ een dynamisch environment allocation omgevingswet/parkeersensoren-voor-een- verkeers- (mobility) dynamisch-verkeersverwijssysteem verwijssysteem (continued) The use of public sector data analytics Table A1. Type of Application (name) Organizations involved Domain application Status Source URL Waterproef/ Municipalities in Friesland, Social services Inspection and Unknown www.hoogeveen.nl/bis/dsresource? waterproof Groningen, Drenthe, SVB, enforcement objectid=d3525063-c977-4300-b159- Belastingdienst d114a3810a04 Project Regionale 130 municipalities Social services Inspection and Unknown TK 2010–2011, 17050, nr. 414 (bijlage) Coördinatiepunten enforcement Fraudebestrijding Noord, Overijssel, Gelderland, Flevoland Schoon schip Municipality of Amsterdam, UWV, Social services Inspection and Unknown TK 2010–2011, 17050, nr. 414 (bijlage) Belastingdienst, SVB enforcement E-Discovery Nationaal Archief, ICTU Data Maintenance Experiment www.nationaalarchief.nl/archiveren/ management nieuws/e-discovery-machine-learning-en-e- mail Kamer van Kamer van Koophandel Economy Inspection and Implemented TK 2010–2011, 17050, nr. 414 (bijlage) Koophandel enforcement DUO DUO Education Inspection and Implemented TK 2010–2011, 17050, nr. 414 (bijlage) enforcement WGA project Gemeente Vlissingen Social services Inspection and Unknown TK 2012–2013, 33 400-XV, nr. 5 Middengebied enforcement Vlissingen Rechtmatigheids- Inlichtingenbureau Social services Inspection and Implemented TK 2012–2013, 33 400-XV, nr. 5; www. controle enforcement inlichtingenbureau.nl/ Inlichtingenbureau Team Darkweb Financial Intelligence Unit Safety/security Investigation Experiment TK 2017–2018, 29 911, nr. 200, p.12 Nederland (FIU-Nederland) (criminal activity) Organized Crime Financial Intelligence Unit Nederland Safety/security Investigation Unknown TK 2017–2018, 29 911, nr. 200, p.11 Field Lab (FIU-Nederland), Nationale Politie, (criminal Openbaar Ministerie, municipalities activity) of Amsterdam and Rotterdam, Belastingdienst, FIOD, Inspectie SZW (continued) TG 15,4 Table A1. Type of Application (name) Organizations involved Domain application Status Source URL SIGMA Nationale Politie, Koninklijke Migration Inspection and Implemented TK 2017–2018, 29 911, nr. 200, p.13; www. Marechaussee, Expertisecentrum enforcement rijksbegroting.nl/2016/verantwoording/ Mensenhandel en Mensensmokkel, jaarverslag,kst233798_13.html þwww. IND, Inspectie SZW, Openbaar rijksictdashboard.nl/projecten/301498 Ministerie DigiDijk Rijkswaterstaat, Physical Maintenance Unknown https://zoek.officielebekendmakingen.nl/ Hoogheemraadschap Delfland environment blg-820189.pdf (water management) Landelijk Meetnet – Physical Inspection and Experiment www.overheid.nl Luchtkwaliteit environment enforcement (air quality) FinPro Openbaar Ministerie, Municipality Safety/security Investigation Experiment www.meetingmoreminds.com/wat-doen-we/ of Rotterdam, Programmabureau (criminal overige-projecten-initiatieven/ NPRZ, Politie activity) Debt Alert CJIB Government Personalization Experiment https://ibestuur.nl/partner-minjenv/kennis- finances tanken-in-vier-pitstops Hansken Politie Safety/security Investigation Implemented https://zoek.officielebekendmakingen.nl/ (criminal blg-849198.pdf activity) De Raffinaderij Politie Safety/security Investigation Unknown https://zoek.officielebekendmakingen.nl/ (criminal blg-849198.pdf activity) MyStatLab Universities Education Personalization Unknown www.overheid.nl Schengen Informatie Politie Safety/security Investigation Implemented www.overheid.nl Systeem II (SIS II) (criminal activity) (continued) The use of public sector data analytics Table A1. Type of Application (name) Organizations involved Domain application Status Source URL Image recognition: IND, TNO Migration Resource Experiment https://socialmediadna.nl/innovatiecongres- 50.000 birth allocation justitie-en-veiligheid/ certificates Logeerfraude Municipalities Safety/security Inspection and Unknown www.wrr.nl/publicaties/working-papers/ (criminal enforcement 2016/04/28/big-data-voor-fraudebestrijding activity) Diagnose Plan en Municipality of Boxmeer Safety/security Inspection and Implemented www.binnenlandsbestuur.nl/sociaal/ Sturingsinstrument (criminal enforcement achtergrond/achtergrond/het- activity) bijstandswonder-van-boxmeer.699608. lynkx Quin-technologie Pandora Intelligence and TNO Safety/security Investigation Unknown https://beveiligingnieuws.nl/nieuws/ (criminal kunstmatige-intelligentie-voorspelt- activity) vluchtroute-criminelen Alert Municipalities of Utrecht, Tilburg, Social services Inspection and Unknown Uitspraak Gerechtsh of Den Haag 28 Nijmegen, Middelburg, Hilversum, enforcement January 2014; ECLI:NL:GHDHA:2014:71 Amersfoort, Leiden, Almere, Groningen, Zwolle, Tiel, Drechtsteden SmartBox UWV, SVB, Belastingdienst, DUO, Safety/security Investigation Experiment https://zoek.officielebekendmakingen.nl/ Justid, CJIB, RDW (criminal blg-814167.pdf activity) LAA - Adreskwaliteit Ministry of BZK Data Inspection and Unknown www.rijksoverheid.nl/binaries/ BRP management enforcement rijksoverheid/documenten/rapporten/2016/ 03/21/rapport-landelijke-aanpak- adreskwaliteit/rapport-landelijke-aanpak- adreskwaliteit.pdf Monitor vitale Municipalities of Putten. Heerde, Physical Maintenance Experiment http://vitalevakantieparken.nl/default.aspx vakantieparken Epe, Harderwijk, Ede, Ermelo, environment Province of Gelderland (public sphere) (continued) TG 15,4 Table A1. Type of Application (name) Organizations involved Domain application Status Source URL AI bij meldingen Municipality of Dordrecht Physical Maintenance Experiment www.smartcitydordrecht.nl/gemeentelijke- openbare ruimte environment projecten/dordrecht-koploper-in-onderzoek- (public sphere) gebruik-artificial-intelligence-ai-bij- meldingen-openbare-ruimte Transparantie VNG Labor market Forecasting Experiment https://depilotstarter.vng.nl/projecten/ informatie over de datagedreven-werken/transparante- arbeidsmarkt informatie-over-de-arbeidsmarkt Themacontrole mbv VNG Social services Inspection and Experiment https://depilotstarter.vng.nl/projecten/ machine learning in enforcement datagedreven-werken/themacontrole-mbv- het sociaal domein machine-learning-het-sociaal-domein Opsporing Municipality of Zaanstad, VNG Safety/security Investigation Experiment https://depilotstarter.vng.nl/projecten/ ondermijning (criminal datagedreven-werken/opsporen- activity) ondermijning-met-data-analyse Brid.ge Municipality of Enschede, Saxion, Physical Personalization Experiment http://techforfuture.nl/onderzoek/afgerond/ Winkelhart Enschede, Twente environment brid-ge/ University, Both Social, eLabbs, (public sphere) Marotura, Present Media, Squal Media, Think Public Thermometer Municipality of Noordoostpolder Physical Maintenance Unknown www.destentor.nl/noordoostpolder/ vandalisme environment thermometer-voor-vandalisme-in- (public sphere) drontena375445d/ Gegevens-vergelijker Municipality of Súdwest-Fryslân Physical Inspection and Experiment https://github.com/EduardWitteveen/ environment enforcement GegevensVergelijker þ www. geoinformatienederland.nl/agenda/gin- bijeenkomst-themagroep-noord-bij-de- gemeente-sudwest-fryslan Proeftuin binnenstad- Municipality of Groningen Physical Forecasting Experiment https://groningen.nieuws.nl/gemeente/ monitoring environment 736506/innovatieve-sensoren-moeten- (mobility) binnenstad-toegankelijker-maken/ (continued) The use of public sector data analytics Table A1. Type of Application (name) Organizations involved Domain application Status Source URL Cold case Politie Safety/security Investigation Experiment www.politie.nl/nieuws/2018/mei/23/00- (criminal nieuwe-technologie-in-oude-politiezaken. activity) html Warme frontoffice Belastingdienst, DUO, CJIB, Raad Social services Personalization Experiment www.slimmernetwerk.nl/2014/05/ voor Rechtsbijstand, bureau WNSP doorbraak-doetank-schuldhulpverlening-in- zicht/ Dashboard Veiligheid Municipality of Tilburg Safety/security Investigation Experiment https://hetccv.nl/onderwerpen/ Tilburg – SpotOn informatiepositie-gemeenten/ voorbeeldgemeenten/gemeente-tilburg/ Verwijsindex VNG, Municipality of Tilburg Social services Personalization Experiment https://hetccv.nl/onderwerpen/ Risicojongeren (youth care) informatiepositie-gemeenten/ voorbeeldgemeenten/gemeente-tilburg/ þ https://vng.nl/files/vng/ 20160223_verwijsindex_risicojongeren_hv. pdf þ https://vng.nl/onderwerpenindex/ jeugd/jeugdhulp/verwijsindex- risicojongeren-vir%20 Big data Pilot Municipality of Apeldoorn Social services Investigation Experiment http://bigdatagemeenten.nl/wp-content/ Apeldoornse Jeugd (youth care) uploads/2017/10/Presentatie-Big-Data- gemeenten_27092017.pdf Scorecard data Municipalities of Enschede, Almelo, Data Maintenance Experiment https://depilotstarter.vng.nl/sites/default/ management Bloemendaal, Heemstede, Loon op management files/project_bestand/ Zand, Sudwest-Fryslan, Borger- overzicht_resultaten_publicatie_0.pptx Odoorn, Leerdam, Terneuzen, Leiden, Venray, Eindhoven, Noordoostpolder, Zoetermeer Geo-data in sociaal Municipality of Uden Social services Personalization Implemented www.digitale-sociale-kaart.nl/uden domein (continued) TG 15,4 Table A1. Type of Application (name) Organizations involved Domain application Status Source URL Website en surfgedrag Municipalities of ’s-Hertogenbosch, Not domain- Personalization Experiment https://depilotstarter.vng.nl/sites/default/ Vught, Utrecht, Veenendaal, specific files/project_bestand/rapport-exploratief- Zeewolde, Veldhoven, Den Haag, onderzoek- Heerlen, Oss, Schijndel digitaledienstverlening20met20huisstijl.pdf Viralcom – Egg Municipality of Dordrecht Physical Resource Experiment www.smartcitydordrecht.nl/projecten- environment allocation innovatiefonds/viralcom-egg (safety on the water) c6Volg Municipalities of Zwijndrecht, Social services Resource Experiment https://depilotstarter.vng.nl/sturen-op- Dordrecht (youth care) allocation resultaten/nulmeting-effectmeting-kosten- sociaal-domein þ https://depilotstarter.vng. nl/sites/default/files/project_bestand/ stapeling_in_het_ sociale_domein_de_100_van_dordrecht.pdf Geniq GGZ instellingen Social services Resource Unknown www.pinkroccade-healthcare.nl/product/ (mental health allocation geniq-ggz/ care) Dashboard Municipality of Zaanstad Social services Inspection and Experiment https://depilotstarter.vng.nl/sites/default/ slachtoffers van (mental health enforcement files/project_bestand/ huiselijk geweld care) monitor_huiselijk_geweld.docx Pronto-app Municipality of Rotterdam/ Physical Personalization Experiment https://ferm-rotterdam.nl/nl/nieuws/ Havenbedrijf Rotterdam environment havenbedrijf-rotterdam-lanceert-pronto- (infrastructure) app-voor-serieuze-efficientieverbetering Artificial inspector Rijkswaterstaat Physical Maintenance Unknown www.fabriek.ai/casus-rijkswaterstaat/ environment (infrastructure) Slimme laadpunten Provinces of Overijssel, Gelderland Physical Resource Experiment www.rtlnieuws.nl/economie/artikel/ environment allocation 4278011/gelderland-en-overijssel-plaatsen- (infrastructure) 2250-slimme-laadpalen (continued) The use of public sector data analytics Table A1. Type of Application (name) Organizations involved Domain application Status Source URL Pilot Jeugdbeleid Municipality of Rotterdam, TNO, Social services Personalization Experiment www.tno.nl/nl/aandachtsgebieden/ Rotterdam Ministry of BZK (youth care) strategische-analyses-beleid/ expertisegroepen/strategy-policy/het-policy- lab-datagedreven-beleidsontwikkeling/ Slimme Camera’s Rijkswaterstaat, Innovatiecentrale, Physical Resource Experiment https://staticresources.rijkswaterstaat.nl/ ViNotion environment allocation binaries/Project%20Slimme%20Camera% (infrastructure) 27s_tcm21-107650.pdf Digitale identificatie VNG Data Resource Experiment www.digitaleoverheid.nl/actielijn/ en verificatie management allocation onderzoek-naar-gebruik-automatische- gezichtsherkenning-en-andere-vormen-van- biometrische-identificatie/ Opsporen ’onbekende’ Municipality of Eindhoven Physical Investigation Experiment https://depilotstarter.vng.nl/projecten/ bedrijven met environment omgevingsdomein/opsporen-onbekende- milieurisico (environmental bedrijven-met-milieurisico-door-data- services) analyse Brandweerradar Brandweer, TNO Safety/security Resource Unknown www.brandweer.nl/media/4153/bom0716- allocation def-lr.pdf Chatbot Wout Politie Oost-Nederland Safety/security Personalization Experiment www.tubantia.nl/regio/politie-oost- nederland-krijgt-hulp-van-chat-robot- wouta3de02a5/ Detectiesysteem Six municipalities Safety/security Inspection and Unknown www.amersfoort.nl/nieuws/sensoren- vuurwerkoverlast enforcement helpen-vuurwerkoverlast-tegen-te-gaan.htm Amsterdam Discovery Municipality of Amsterdam, app- Physical Personalization Implemented www.computable.nl/artikel/nieuws/ Challenge-app developer environment overheid/6161734/250449/app-moet- (tourism) toeristen-verspreiden-over-amsterdam.html Rijenradar City marketing organization Physical Personalization Experiment www.computable.nl/artikel/nieuws/digital- Amsterdam Marketing, environment innovation/6002561/250449/rijenradar- Municipality of Amsterdam (tourism) toont-wachtrijen-amsterdamse-musea.html (continued) TG 15,4 Table A1. Type of Application (name) Organizations involved Domain application Status Source URL Smart Roof 2.0 Municipality of Amsterdam, Physical Maintenance Experiment https://amror.nl/smart-roof-2-0/ Waternet, Drain Products, Aedes environment Real Estate, KWR Water Cycle (water Research Institute, Marineterrein management) Amsterdam Kunstmatige Rechtbank Oost-Brabant Judiciary Resource Experiment www.rijksoverheid.nl/binaries/ Intelligentie allocation rijksoverheid/documenten/kamerstukken/ Kennissysteem voor 2018/12/19/ek-artificiele-intelligentie-en- de rechtspraak algoritmen-in-de-rechtspleging/ek- artificiele-intelligentie-en-algoritmen-in-de- rechtspleging.pdf Watson RVO RVO, Jibes Physical Personalization Experiment Eindverslag Pilot Watson RVO environment Digital we Manifestgroep, Municipality of Not domain- Personalization Experiment Presentatie bestuurlijke bijeenkomst 19 nov, Amsterdam, SIVI, Bvolve, specific 2018 Innovalor Waterbeheer SR Tauw, HAL24k Physical Maintenance Experiment Data challange inzending Tauw_Hal24k environment (water management) Statistiek van CBS, Kadaster Physical Forecasting Experiment www.kadaster.nl/-/cbs-en-kadaster-maken- nieuwbouw- environment statistiek-van-nieuwbouwtransacties? transacties redirect=%2Fweb%2Fkadaster.nl% 2Fterzake Slimme sensor Municipality of Rotterdam Physical Resource Implemented www.ad.nl/rotterdam/afvalcontainers- afvalcontainers environment allocation voorzien-van-slimme-sensoraf64633b/ Opsporing ABNAmro, University of Social services Inspection and Implemented https://nos.nl/artikel/2273390- arbeidsuitbuiting Amsterdam, Inspectie SZW enforcement mensenhandel-opgespoord-via- (mensenhandel) bankgegevens.html; www.bnr.nl/podcast/ internet-vandaag/10370844/abn-amro-jaagt- met-algoritmes-op-mensenhandel About the authors The use of Dr Anne Fleur van Veenstra is a Senior Scientist in ICT Policy at TNO’s Strategy and Policy public sector Department. Her research focuses on public sector data analytics, data-driven policy-making and data analytics public sector innovation. She has published on data-driven policy, digital government, open data and public value management. Before joining TNO, she worked as a Researcher in the field of electronic government at the Department of Technology, Policy and Management of Delft University of Technology, where she wrote her PhD thesis, entitled “IT-induced public sector transformation.” Dr Francisca Grommé is an Assistant Professor at the Organisational Dynamics of the Digital Society. Her research focuses on digitalization in relation to organizational knowledge practices: how digitalization changes the way data are collected, analyzed and verified by the state and in the digital economy. She has published about topics such as surveillance, identity, nationhood, digital work and citizenship. She previously worked at the Netherlands Organisation for Applied Scientific Research (TNO) and Goldsmiths, University of London. Her PhD thesis was written at the University of Amsterdam (AISSR) and concerns the role of technology experiments in policing. Francisca Grommé is the corresponding author and can be contacted at: gromme@essb.eur.nl Somayeh Djafari, LLM is a Senior Lawyer at the Ministry of Health, Welfare and Sport. She is also a PhD Candidate at the Law Faculty of the University of Leiden, where she is working on balancing the interests of the community as a whole and the right of the individuals affected by the legislation to respect for their private life and home in embedding the algorithm(s) by administrative authorities in the Netherlands. This article is written in a personal capacity. For instructions on how to order reprints of this article, please visit our website: www.emeraldgrouppublishing.com/licensing/reprints.htm Or contact us for further details: permissions@emeraldinsight.com

Journal

Transforming Government: People, Process and PolicyEmerald Publishing

Published: Nov 30, 2021

Keywords: Data analytics; Big data; Open data; Artificial intelligence; Evidence-based policy-making; Data-driven policy-making

References