Open Data Policy in Estonia: An Opportunity for Innovation in the Media Industry?

(1)

University of Amsterdam Faculty of Humanities

Media Studies

Open Data Policy in Estonia:

An Opportunity for Innovation in the Media Industry?

Hanna Jemmer 12221694 Kaupmehe 4-21, Tallinn, Estonia hanna.jemmer@gmail.com +372 5345 2122 Date of completion: 23.06.2020

Supervisor: Stefania Milan Second reader: Stijn Peeters

(2)

1

Introduction

This paper explores the Estonian open data portal opendata.riik.ee and its use by legacy media organisations. More specifically, potential benefits of open data for the mainstream media and the related barriers and drivers are studied. Open data as a topic emerges in research regarding open government and transparency (Toots, McBride and Kalvet 2017), public policy studies (Dawes, Vidiasova and Parkhimovich 2016), business studies (Jetzek, Avital and Bjorn-Andersen 2014), media and journalism studies (Radchenko and Sakoyan 2014), and media activism (Milan 2017). The paper mainly draws on media and journalism studies, public policy research and business studies.

There has been some research around the governmental and non-governmental initiatives concerning open data that among other countries focus on Estonia (McBride et al 2018; Toots et al 2017) but overall there are very few studies on how private sector uses open data. Some examples include Jetzek et al (2014) and Susha et al (2015). Estonia is a relevant case to study as the public sector through collects and connects a lot of governmental data digitally. Additionally, many public organisations have the obligation by law to collect and publish data in an accessible manner. Therefore, the established system of e-services has created sources of data. As so many services are provided online, it is therefore fair to assume that the citizens are used to the system as well. Theoretically, this provides a fertile ground to explore the situation regarding the provision and use of open data. However, as it happens, Estonia has for many years lagged behind in rankings regarding open data (McBride et al 2018). This creates a situation where there are well-functioning digital services but citizens, non-governmental watchdogs, companies and the media do not find open data useful or applicable for innovation. This apparent contradiction makes the Estonian open data portal and its use by legacy media organisations relevant. With that in mind, the research question is how the legacy media is using open data to provide their services in Estonia.

Turning to the relevance of the user side in exploring open data, legacy media makes a compelling case. By legacy or mainstream media, I mean newspapers, television stations and magazines that have been around for decades. Legacy media is a relevant example to study as it aims to balance private and public interests much like open data itself that can offer both public and private value. Further, legacy media is often positioned to oppose new media. New media as a term explained by the Oxford Dictionary stands for a means of mass communication through digital means. Examples include platforms/ services like Facebook, Instagram, Google, Netflix, Spotify etc. that offer content via digital structures. Mainstream media is pitted against new media because for the past decades, advertisement business that has been crucial for legacy media to exist, has been turning towards new media businesses due to more advanced segmentation and targeting technologies allowing for more personalised ads and content. According to Kõuts-Klemm et al (2019) this change has taken place in Estonia as well where profits from ad sales have declined significantly as Google and Facebook have taken over more and more advertisement spaces (96). In addition to creating private benefit (i.e. profits through ad sales), legacy media also has a responsibility to provide public benefit. The public benefit or value that legacy media provides is the creation of a public sphere, a concept based on Habermas but extended by Caplan & boyd (2016). Legacy media allows for the formation of public opinion, it also acts as a watchdog revealing imbalances in power. According to Deuze (2005) there are codes of conduct that journalists follow to provide the public with information and critique about political and social events. This distinguishes legacy media organisations from new media companies that avoid public scrutiny and regulation by claiming the role of a technology company instead of identifying as media companies. With that new media companies escape regulation and responsibility- a phenomenon witnessed by many authors including Zuboff (2015) and van Dijck & Poell & De Waal (2018).

(4)

3

Mazzucato’s (2018) concept of value creation can be applied here as well. Namely, platforms seem to extract value via existing content and actions of users while providing very little (i.e. opaque data collection and sharing practices) in return. Legacy media has had a somewhat different role as it has provided topical information/ content to the public by using its own human resources. Therefore, as public value creation via legacy media is relevant, then it is relevant to explore its opportunities in employing open data to provide both public and private benefits. Open data stands for data produced or commissioned by public bodies or government controlled entities which is made accessible, can be freely used, reused and redistributed by anyone. In other words, it is governmental data that acts as a public resource and is therefore made available for the public (non-governmental organisations, media, private companies etc.) to benefit as well. Therefore, as open data ideally allows both business (i.e. automation of processes) and content related (i.e. timeliness) benefits, it is necessary to explore the potential benefits, barriers and drivers related to open data use by legacy media. Having highlighted the existing knowledge gap and how the paper aims to contribute, I now present the structure for the paper.

At first, relevant literature is presented aiming to provide a context to the study and the discussion. As it has been explored from different perspectives- public policy, media and journalism, digital activism and innovation & business studies, these approaches are introduced to provide a general framework. Secondly, methods are explained. In summary, the paper makes use of content analysis and semi-structured interviews. First, content analysis of two relevant policy documents is used - the Estonian Digital Agenda 2020 and the Green Paper on Open Data are studied. It helps to establish the government’s priorities and the legacy media’s relevance within those. Furthermore, the governmental open data working group meeting minutes are explored for answers regarding most prominent issues and developments of the open data portal. Additionally, six interviews were conducted to offer more precise input on the specifics of the legacy media sector and its use of open data in view of potential benefits and the existing barriers and drivers. The interviews were conducted with the supplier side (the government and the non-governmental organisation that administers the portal) and the user side (a scholar who has worked studying open data, a journalist using open data, a media innovation expert and a data analyst from a media organisation).

Then follow two chapters on findings where the first gives a more general overview of the open data portal and its more relevant developments since 2014 by reflecting on the content from the two mentioned policy documents. It unpacks the governmental attitude towards open data provision more generally. The next chapter discusses findings from the legacy media view and focuses on sector specific benefits, barriers and drivers. This is followed by a chapter discussing the findings and potential for innovation via open data. Lastly, the paper is summarised and further research needs are suggested in the conclusion.

1. Literature review

To map how legacy media in Estonia approaches the opportunities associated with the use of open data, an exploration of existing literature to build a general framework is necessary. Open data as a topic emerges in research regarding open government and transparency (Toots, McBride and Kalvet 2017), public policy studies (Dawes, Vidiasova and Parkhimovich 2016), business studies (Jetzek, Avital and Bjorn-Andersen 2014), media and journalism studies (Radchenko and Sakoyan 2014), media activism (Milan 2017). The paper mainly draws on media and journalism studies, public policy research and business studies. First, the majority of open data research deals with the benefits, barriers and drivers of open data. The first section of the literature overview therefore highlights the more general views on open data and the possibilities related to it. The second part of the literature review summarises how enterprises have been positioned within the open data ecosystem. It highlights issues such as value creation, both economic and social, and potential innovation opportunities. The third part introduces the research done on open data in Estonia, highlighting the current situation and some research gaps.

(5)

4

1.1. Open data benefits, barriers and drivers

Open data is defined based on the European Data Portal’s definition from their report on The Economic Benefits of Open Data (2017). It stands for “data produced or commissioned by public bodies or government controlled entities which is made accessible, can be freely used, reused and redistributed by anyone“. Open data was declared by the European Commission’s Vice President Neelie Kroes the new gold in 2011 (European Commission 2). Also in the European Union (EU), public sector is one of the most data intensive sectors as it holds vast amounts of public sector information which has led to respective legislation in the EU starting from the Public Sector Information (PSI) directive in 2003, renewed in 2013 and lastly amended in 2019. In the 2013 Directive the scope of application was expanded to include cultural organisations, costs for access to data were limited, and it invited governments to introduce machine-readability of data. The changes in the 2019 version aim to improve free re-use of open data, the Directive identifies high-value datasets (i.e. geospatial, statistics, mobility etc.), and encourages more real-time data via Application Programming Interfaces (APIs). The shift from opening up government data for making public sector information more accessible to achieve transparency has been complemented by adding a focus on potential economic gain. The total economic value is expected to reach 194 billion euros by 2030 (European Commission).

In academic literature developments on open data (or public sector information)1_{have been}

discussed more actively since 2011-2012 when Open Data Initiatives and Open Government Partnerships had been launched in several countries. Open data initiatives stand for governments taking up responsibilities to publish data openly (The World Bank), and Open Government Partnership is a multilateral initiative signifying the coming together of the government and the civil society to promote open data (Open Government Partnership). Most literature focuses on the barriers or drivers for open data use (Barry & Bannister 2014; Conradie & Choenni 2014; Janssen et al. 2012) and its benefits. These studies rely on case studies and analyse open data issues on different levels of government. As Barry and Bannister (2014) show, there are many categorisations of barriers in the literature (131). One is a data centred category where usability, quality, availability have been named as issues. This category has to do with the technical side where data itself has to possess certain characteristics to be useful. These characteristics relate to the provision principles by public organisations where often there is a lack of rules to dictate the data format and the overall quality.

The second category of barriers has to do with the provision or release of open data. This means that data can lead to fear of false conclusions where the representability and complexity of data is unclear. Also, opaque ownership of data can be problematic when a service based on different providers of open data ends up as a (policy) solution where different practices of data collection and provision cannot be tracked back to identify biases (Conradie and Choenni 513). Another concern originates from the financial side, mainly the question whether it is ethical to ask a fee for providing data (Conradie and Choenni 513). Therefore, the release of data can present both barriers and drivers depending on the specific data and its provision principles.

The third category is the use of data that relates to the previous category of releasing data where data from different sources when used can clash in terms of format or quality. In other words the data is not interoperable and can cause fragmented results (Barry and Bannister 131-133). The category of use of data also contains the skills necessary for handling data and creating new applications based on that data (Susha, Grönlund and Janssen 29). Therefore, in order to create something useful and comprehensive, data analysis and other computational skills are needed. If there is no activist or business community with technical skills, the provision and interoperability of data are also irrelevant as there are only limited people with skills. Zhang et al. (2005) in turn

1_{I use the terms interchangeably}

(6)

5

emphasise the importance of varying perspectives of stakeholders (135) which means that barriers and drivers depend on the participant, mainly their sector and organisational/ business goals. The focus on barriers and drivers reflects a general agreement about the benefits of open data. Besides the financial benefits important for the European Commission and the EU Member States in their approach to open data, authors invested in researching open data claim other benefits. Dawes et al (2016) name citizen engagement which stands for citizens engaging with open data to compare, combine and produce data based applications for public use and scrutiny over policy makers (15); Janssen et al list greater transparency and accountability that is echoed throughout the literature. Janssen et al continue with suggesting that transparency and accountability lead to improved policy making and increase citizen satisfaction (Janssen, Charalabidis and Zuiderwijk 260). However, there are sceptics like Morozov (2013) and Hale (2008). Hale suggests that transparency can be harmful to good governance mainly because it can encourage defensive behaviour by the public sector and with that decrease efficiency (Barry and Bannister 135). Morozov in turn dismisses efficiency but states improved decision making as the main goal for using open data (86). He adds that the mentality emphasising openness and transparency in view of open data should not be praised for their own sake but that these principles should be instrumental (Morozov 86). That means transparency should lead to improved decision making and not be an end in itself.

In the discussion around what open data should achieve, the main question seems to be about what is appreciated in a society- efficiency in decision making, transparency, accountability, economic gains or justified decision making. The most interested parties are activists from NGOs, public service providers, local and national governments, academics and the start-up community. It is therefore a question of political and societal priorities and shows that different benefits can be seen important when dealing with open data.

Jetzek et al (2014) present a useful division to position the economically beneficial use of open data. They divide open data mechanisms into four categories where the first stands for transparency- mechanisms to reveal relevant information that is being generated (106). The second deals with participation mechanisms that enable public participation in government (Jetzek, Avital and Bjorn-Andersen 106). The third is efficiency mechanisms that seek to improve resource utilisation (Jetzek, Avital and Bjorn-Andersen 107). The fourth is the category where to position the media sector data use that I am interested in- that is innovation mechanisms. These are designed to generate new or improved products and services that can transform markets and industries by creating new value (Jetzek, Avital and Bjorn-Andersen 107). Therefore, in view of analysing the media sector, the social and the economic value potentially generated by the use of open data fits well into this framework as a starting point as other categories are more directly related to the functioning of the public sector. Questions on accountability and transparency have dominated the open data related literature, therefore the issue of economic benefits via open data remains underexplored. I turn to summarising the research about the economic benefits via open data, and later in the empirical part aim to contribute to the debate by exploring the situation in the media sector in Estonia.

1.2. Use of open data in service innovation

Co-creation of public services based on open data by involving different stakeholders is mentioned by Toots, McBride and Kalvet (102). However, here the goal is to outsource public services not perhaps create profitable business models. So besides promoting transparency, accountability and public engagement, innovation by private companies should increase entrepreneurial activity, growth of start-ups, and better business services (Susha, Grönlund and Janssen 20). According to Susha et al (2015), the academia has understudied how the private sector uses open data (20) despite confirmations by for example the World Bank, the European Commission, and the Organisation for Economic Co-operation and Development (OECD) about the economic benefits of open data

(7)

6

based innovation by private companies. Still, there have been efforts by authors such as Dawes et al (2016), Jetzek et al (2014), Janssen et al (2012) and Susha et al (2015) to navigate this field by approaching this from the public policy and business innovation angles.

Janssen et al (2017) and Susha et al (2015) suggest that service innovation on the basis of open data is limited and businesses lack the knowledge on how to make use of open data for economic benefit. Also, as Susha et al suggest, motivations for different companies are contingent on whether they are start-ups or more established companies, moreover environments in different countries and approaches in different sectors vary (Susha, Grönlund and Janssen 32-33). With that they confirm Zhang’s (2005) proposition that drivers and barriers depend on the participant. Further, also sensed barriers and drivers vary within the same sector. For example, more established technology companies do not find skills for accessing and analysing open data important while start-ups name this as an issue more often (Susha, Grönlund and Janssen 29). For data journalism, Stoneman (2015) echoes a similar concern, saying that data journalism also needs a certain skillset (Stoneman 1) where without the skills, there can be no meaningful data journalism.

Based on the literature (Conradie and Choenni; Toots, McBride and Kalvet), it seems there is unspoken agreement that when businesses use open data, they should provide a service that benefits the public not just the companies themselves. This seems debatable as the European Commission itself has calculated the potential of open data innovation in billions of euros. If public services are concerned, the measurements should perhaps be in savings, effectiveness and/ or convenience, and not in profits. It begs the question, should the focus in view of open data always be on public benefit and public value. As Susha et al (2015) state, for businesses, there often exists a strictly cost-benefit view when the potential of open data is considered for innovation, except when the company is a social enterprise and thus the goal is positive influence on the community (21, 29). Barry & Bannister (2014) also confirm that benefits and barriers depend on the participant (130). This suggests that different companies have various motivations to innovate via open data, which means that narrowing the research focus is necessary to provide more concrete answers on the barriers and drivers of open data. Radchenko and Sakoyan note that for the private sector and for journalism, it is business opportunities that induce the use of open data (Radchenko and Sakoyan 48). They add that the lack of viable business models for use of open data undermines the use of open data (48). Therefore, support from the governments is expected (Radchenko and Sakoyan 48) to pinpoint the opportunities of open data concerning different sectors and businesses. Legacy media or mainstream media stands for newspapers, television stations and magazines that have mostly been around for decades. These companies represent a relevant case as their motivations are two-fold. In that sense they resemble open data that can be used for both private and public benefit. On the one hand, legacy media companies operate as regular businesses with the aim of making profit. On the other hand, their function is to provide information to the public in an objective, neutral and impartial manner (Deuze 447). Governments themselves have formally recognised the role of media informing the citizens and with that advancing democracy (Caplan and boyd 3). Farrington adds another aspect saying that the news media should gather information from different experts to inform the public as well as possible because not everyone has the time and the expertise to individually inform themselves on every little detail in the world (Farrington). Legacy media therefore often acts as an informer and a watchdog. With that in mind, Hale (in Barry & Bannister 135) notes the interesting position of media aiming to increase transparency and openness of the public sector. Hale says that media’s demands for further transparency in public affairs can lead to blame avoidance and defensive behaviour that results in the opposite effect (Barry and Bannister 135). To clarify, public authorities can become fearful of criticism and as a result avoid being open in their opinions. Achieving and maintaining trust is one priority of the public sector and therefore complete openness can be seen as a risk.

(8)

7

Another layer to the public vs private interest debate can be added by using Mazzucato’s conception of value (2018). As said, legacy media has largely two responsibilities, one being economically sustainable and the other is providing the public with impartial and objective information. Provision of information in a responsible and impartial manner can be treated as creating value for the society- an educational, mind-broadening value, I would argue. The opposite, according to Mazzucato is extracting value which means re-arranging already existing assets and asking a price for it for private benefit (6). In other words, value extraction stands for wealth creation not value creation (6). The data economy2_{and the ensuing platform society}3_{(van Dijck,}

Poell and De Waal 2018) could be seen as a value extraction endeavour and the Silicon Valley has been one of the driving forces directing societies towards data economies. This data collection centred model for businesses has become dominant and has clear downsides.

The Silicon Valley companies have presented themselves from the 1980s onwards as the creative do-gooders when it comes to innovation and job creation. The terms like “sharing economy” and “smart growth” seemingly represent this positive approach used by Facebook and Google (Mazzucato 189-90). The reality is a bit more nuanced. This Silicon Valley mentality to offer ‘empty’ platforms for user interaction and content creation represents a supply-demand model to connect the two sides of the market offering just the infrastructure. The private gain comes from collecting and selling personal data (Mazzucato 215-16). However, these data intensive platforms avoid responsibility- they have elaborate tax schemes, face no accountability for the content, and their data collection practices are opaque. Though in their narrative, these companies claim to create public value- they claim to connect communities amongst themselves and with information, and with that stress that the governments are ineffective and useless in those tasks (van Dijck, Poell and De Waal 23). This has led to a broader mentality that the governments should have a hands-off approach to allow for businesses arrange societal affairs. However, in short, when it comes to paying taxes (sharing the wealth) or revealing the setup of their data collection practices, there seems to be no public accountability.

The Silicon Valley narrative stating that governments should have a hands off approach to innovation as their participation hinders private sector from advancing is overturned by Mazzucato (2018). She brings examples how governments have been behind investments into riskier technological projects such as Arpanet which was later a basis for many innovations by the private sector (194). Therefore, the public sector has a role in both ideologically and practically to support innovative ideas as its position is more durable to risk and change. Innovation as a collective process requires different stakeholders to be involved and active as participants. This means, exploring both public and private sector perspectives on the benefits of open data in view of innovation is equally important in this paper and also more generally in terms of policies.

Internet based media in its early phases was praised for its democratic potential- giving a voice to everyone- but it has not stayed free from surveillance and censorship (Caplan and boyd 3). The debates take place on platforms that operate with concealed algorithms and limiting affordances and that has led to high personalisation creating filter bubbles and polarisation (Caplan and boyd 3). Algorithms in the media industry have come to play an increasingly big role to navigate the audience’s wishes (Napoli 1). The media is trying to navigate the fragmented market by using big data and algorithms for mainly two purposes- as a demand predictor and as a content creator (Napoli 1). However, in view of legacy media especially in smaller countries, these business

2_{According to the European Commission’s (EC) definition data economy is A data economy is a global digital}

ecosystem in which data is gathered, organized, and exchanged by a network of vendors for the purpose of deriving value from the accumulated information (from the 2017 EC communication on Building a European Data

Economy, Digital Single Market).

3_{According to the authors, one of the definitions of platform society is a society in which social and economic}

traffic is increasingly channeled by an (overwhelmingly corporate) global online platform ecosystem that is driven by algorithms and fueled by data.

(9)

8

opportunities were discovered too late and therefore there are limited capabilities (Farrington) which results in diminished relevance of the content for the public and reduced revenues from advertising. Therefore, in view of this paper, it will be interesting to position legacy media’s approach and the Estonian government’s attitude towards innovation via open data into that framework and see if and how the aim of creating public value balances the goal to gain private benefit.

Open data acts as a source of information and therefore legacy media with its dual motivations represents a fitting example to study barriers and drivers to open data innovation. As Radchenko and Sakoyan state, using open data for data journalism is seen as a way to provide an advantage to the legacy media companies as their influence due to the rise of social networks as an information provider has diminished significantly (Radchenko and Sakoyan 47).

Another way how open data can act as a source of higher competitiveness for national legacy media is because of data’s locality. Odell (2019) brings an elegant parallel to how local information can have high significance. She speaks of a teletype machine which was a machine installed in several locations in California in the 1970s and after receiving calls from locals it printed these community oriented messages on a bulletin board that could be viewed by people passing these boards near shops or cultural organisations (Odell 166). It did not organise data as it is done by Facebook and Google by attaching categories to a user to feed them content based on these categorisations but everyone in the locality could access locally relevant data equally (Odell 169). All that was posted would be seen by everyone. Therefore, there was no filter bubble created by the technology as per the content. In the context of this paper, open data could serve a similar function- that people belonging to a certain community can access information relevant to their locality and the legacy media sector can be an intermediary as it has been historically while keeping in mind its public function. It appears very fitting.

The third possibility for innovation leading to potential increased competitiveness relates to Odell’s (2019) description on how exhausting and superficial the consumption of news has become due to the news feed affordances applied by Facebook and Google that only list a headline and a few sentences as a news story (93). This is due to the fast 24-hour news cycle where new information is presented to capture the reader’s attention over and over again. That however leads to superficial knowledge of topics inducing the aforementioned filter bubbles where people are pushed into. Instead, the media can offer high quality slow journalism by attending to an attentive and loyal public that perhaps pays for content. That however means reconfiguring the affordances of the medium in a way that the readers appreciate and pay attention to long form and more detailed content as happily as to the fast paced news feed. There is a conundrum though, will the affordances of slow media reprogram the reader or will it require other types of reprogramming to help the public slow down and appreciate thorough media content. Transmedia journalism, meaning event reporting combined with compelling storytelling could create enough attention to turn to more sophisticated media content.

In Estonia as elsewhere global data collection giants are putting local media out of business (Kõuts-Klemm, R. et al 96). With its reputation as a digital state where e-government services host vast amounts of data, Estonia with its modest results in open data rankings presents a relevant case study. To reiterate, the legacy media sector is losing in profits to global new media giants like Google and Facebook but on the flip side the government collected data is vast and could therefore be used for media sector innovation purposes. Some of the options for innovation via open data based journalism were sketched. The question however is, can this increase the competitive edge and what the associated barriers and drivers are. Uncovering the barriers and drivers is the first step in the process addressed in this paper. This is why I turn to presenting an overview of research that has concentrated on use of open data in Estonia.

(10)

9

1.3. Open government data in Estonia

Estonia ranks high in e-government services rankings (Plantera). 99 per cent of state services are online, altogether there are 2773 services that can be used via the X-road infrastructure connecting different databases and services with citizens (e-Estonia). The services range from e-health to internet voting, from declaring taxes online to parking via mobile phones. Furthermore, 99 per cent of Estonian residents have an electronic identity card necessary for using these services. For start-up friendliness, Estonia ranks 1st_{in 2018 (start-up friendliness Index Venture), for internet}

freedom it ranks 2nd_{(Freedom House) for entrepreneurial activity Estonia ranked 1}st_{according to}

the World Economic Foundation (e-Estonia). Additionally, Estonia is known for high levels of transparency (McBride et al 614). This means the digital penetration of the society is very high and in combination with high transparency it seems as a country destined for highly innovative digital solutions in the private and the non-governmental sectors. There exist electronic databases with vast amounts of public sector information that can potentially be used in a secure and anonymised manner as open data for creating innovative solutions in the media sector.

Consequently, Estonia represents an interesting case in view of open data provision and use. Surprisingly, as McBride et al (2018) and Kassen (2019) suggest, Estonia is a laggard when it comes open government data (OGD) rankings. The reasons range from the lack of knowledge of OGD, high costs, lack of resources, unclear business cases with unknown payoffs, data quality (619-20). One of the more curious reasons is the well-established e-governance service through the X-road data exchange layer which interconnects government information systems and databases where both citizens and authorised government personnel can access and view information concerning public services provision (McBride et al 618). This suggests that information flows well within the system and people have adequate access to it, so open data services which aim to increase flow of public information become redundant to an extent. Authors such as Toots et al (2017), McBride et al (2018) and Kassen (2019) have discussed the dynamics of open data in Estonia but have up to now focused on transparency, participatory and collaborative aspects of open government data. Therefore, public benefit and public service creation has been in focus but only generally. As Toots et al (2017) state, lack of viable business models decrease the potential of open data innovation (107) but this is something that has not been looked at in detail. If open government data is not used for improving public information flows as allegedly the flow is good enough, then looking at the business opportunities and the innovation potential for entrepreneurs seems necessary as it is the missing piece in the open data ecosystem research. Based on Zhang et al (2005) and Barry and Bannister (2014) who state that benefits, barriers and drivers depend on the participant, it will be possible map the existing situation and set the scene for improvements by exploring potential benefits and the barriers. This means that choosing the use of Estonian open data portal (opendata.riik.ee) by legacy media companies as a case study should highlight the current situation, its potential benefits, and more specific drivers and barriers to the use of open data for innovation. With that, the paper aims at contributing to research on open data based innovation in the private sector.

Based on the existing research which has highlighted gaps when it comes to open data based innovation in the private sector, I have formed one main research question and two sub-questions. 1. How is the legacy media using open data to provide their services in Estonia?

1.1. What are the drivers and barriers for open data use in the media sector in Estonia? 1.2. What are the perceived benefits by the Estonian government and the media sector in view

(11)

10

2. Methodology

In this chapter I present the methods used to answer the main research question and the sub-questions. I first explain why case study as a method is suitable. As case studies combine different tools, I then go on to argue why content analysis in combination with semi-structured interviews of experts is a suitable approach. Lastly, I explain how content analysis and semi-structured interviews ensure saturation of collected data. As a framework I present how the case study relates to the open data ecosystems model stemming from the sociotechnical studies theory.

2.1. Qualitative research via case studies

The phenomenon of open data infrastructure and the multiplicity of stakeholders involved in the decision making concerning open data calls for a qualitative research approach that allows for exploring complex phenomena. Both stakeholders and the procedures can be seen as uncontrolled and complex variables. Therefore, as expressed by Dyson and Genishi (2015), collecting similar types of information from different sources can be greatly beneficial in uncovering case study specific understandings that can be extended to a general understanding of a phenomenon (Compton-Lilly 57), in this case the use of open data by the private sector for financial and public benefit. Moreover, as Yin (2018) states, case study presents a more flexible method and is suitable when the main research questions are “how” or “why” questions, there is little or no control over behavioural events, and the focus of study is a contemporary (as opposed to entirely historical) phenomenon—a “case (Yin 45).” This definition fits the aim of this paper as I am dealing with a ‘what’ question and a complex set of variables in respect to exploring the Estonian open data portal. Estonia is a relevant case as the government has invested a lot into creating and improving digital services for the convenience of the citizens and in return the user side has to a large extent embraced these possibilities. However, as said use of open data seems incompatible with these developments and therefore it is worth studying the reasons by using a sample case, that is how legacy media as a user regards open data benefits and the related barriers and drivers. Having established initial case boundaries, it is important to note that the current paper is an exploratory single case study and the main research question- how is the legacy media using open data to provide their services in Estonia- presents a good grounding for using case study as a method. This method allows to explore a phenomenon from different and relevant perspectives and within it combine other more concrete methods or tools, i.e. content analysis and interviews.

The approach of case studies has also been used by many afore-mentioned academic works, either on the local (Conradie and Choenni), governmental (McBride et al) or business (Jetzek, Avital and Bjorn-Andersen) level. While Conradie & Choenni (2014) have used participatory action research (PAR) and Janssen et al (2012); and Barry and Bannister (2014) more traditional semi-structured group interviews, then Toots et al (2017) and Kassen (2019) study policy documents. Therefore, case study as a method offers more depth and detail on the open data innovation possibilities in the legacy media sector by employing a combination of studying governmental strategy documents and conducting interviews which helps to highlight more concrete benefits, barriers and drivers. In other words, the current paper combines content analysis and interviews as more specific methods. Study of governmental documents on the topic of open data, and semi-structured interviews with experts with more detailed knowledge on the various aspects will give a fuller picture of the situation and hopefully allow for more general conclusions than the current case study analysis in itself.

2.2. Content analysis

According to Bengtsson (2016), content analysis allows to organise and interpret a meaning from the data (the documents) and draw justified conclusions (Bengtsson 8). What is more, it is not just the counting of certain instances but connecting it to the surrounding environment and context (Bengtsson 9) that contributes to creating new knowledge. Therefore, the findings from analysing

(12)

11

the content of policy documents can be reflected upon the existing studies summarised in the previous chapter (literature review). In short, placing current findings into a dialogue with existing literature should create a more comprehensive picture and from that more general conclusions can be drawn. Moreover, adding interviews as a method (described later) will offer more detailed insights that contribute to the better understanding of the research object and its context.

To get a better sense of the opportunities and potential benefits of using open data for innovative media services, more strategic policy documents are chosen as study objects. The documentation around the development of the open data portal is a vital step, therefore the Estonian Government’s Green Paper on Open Data with a mentioned significance on opendata.riik.ee portal deserves scrutiny. Another strategic document dealing with open data developments and possibilities is the Digital Agenda 2020 for Estonia. The relevance of the Green Paper is that it is the only existing policy document on open data in Estonia and the relevance of the other is that it is a more general development document that gives an overall sense of the importance of open data among all topics concerning governmental priorities on information society. Further, there is a working group composed of public sector, NGO and academic representatives that regularly discusses the most relevant issues concerning open data developments. If policy documents lack detail, the minutes of that working group can also be analysed to evaluate to what extent and how the private sector and especially the media sector is represented in the discussions.

When analysing the policy documents, the main questions are whether benefits, barriers and drivers are covered and to what extent. Moreover, is the focus on business opportunities, especially in relation to the media sector present and in how much detail. The same can be deduced from the meeting minutes, more specifically, what the main topics under discussion are, whether business opportunities for the media sector are scrutinised and whether the representatives of the private sector and more specifically the media sector are present. Their opinions on the benefits, barriers and drivers within those documents will give a general sense of the attitudes towards the open data portal. With that, I have set boundaries to what documents are analysed and what information is sought to answer the more general research question.

2.3. Semi-structured Interviews

The other method – interviewing, more specifically, semi-structured interviews will, as said, provide another layer to the analysis. This is done to avoid potential bias and far-fetched conclusions and to offer more relevant data that contributes to the more comprehensive handling of the subject. As Roulston defines it, a semi-structured interview is “a dynamic exchange of ideas based on researchers’ open-ended questions or areas of interest with probes that are designed to elicit details and explanations” (Trainor 126). In short, there are some structured and open ended questions to allow for follow-up questions and comments to uncover new relevant information on the topic. Further, interviews are seen as a complementary tool for data collection as qualitative research aims for a complex and multi-layered view of the research object (Trainor 126) confirming the need for content analysis of policy documents.

To clarify, the sampling strategy corresponds to several criteria set out by Trainor (2013). One of the criteria is that a small number of interviews suit qualitative studies better to reach the saturation point. In other words, the chosen interviewees should help to answer the research question sufficiently (127-8). Saturation is ensured by an informant interview with a scholar who has recently studied open data related policies in Estonia. The interviewees are therefore chosen based on their expertise on open data and innovation either from the governmental, administrative position or from the perspective of media practitioners. This responds to Trainor’s further criteria that the interviews should provide breadth and depth (127). Therefore, different perspectives of the phenomenon under scrutiny will be explored by studying texts (policy documents) and interviewing different stakeholders. This method according to Heath (2001) is a qualitative social science method to reduce error and biases through using multiple measures (15901). This will produce a

(13)

12

more complex understanding of how legacy media positions in the use of open data for its services. So while content analysis will be the basis to assess how much relevance open data, private sector innovation via open data, and media sector related open data instances have, the interviews complement the picture.

As to the sampling of the interviewees, two key members of the working group will be interviewed for more in-depth information that are not set out in the formal documents concerning open data. There are two main members in the working group who have been present at the meetings. Therefore, one interviewee is a representative from the public sector. The other represents the non-governmental organisation (NGO) Open Knowledge Estonia (OKE). The former will provide a governmental perspective on the barriers, drivers, and benefits of the open data portal more generally and also in view of the media sector. The latter acts as an overall coordinator of open data related practices and their expertise lies in open data related principles such as accountability, openness, and transparency. OKE as an organisation stands for sharing knowledge openly which explains its role as a coordinator of open data related issues. Its mission includes advocating for sharing non-personal data for transparency and innovation purposes (Open Knowledge Foundation Estonia).

As the study concerns directly the private sector, specifically the legacy media, then a data journalist from one of the biggest daily newspapers was identified and approached for an interview. The respective expertise lies in the first hand use of open data for pieces of data journalism. As open data possibilities concern not just the content (as in data journalism) but also profitability through innovative business models, then two representatives dealing with media innovation and business development were included. This is useful to cover the potential benefits and barriers of open data from the business perspective. In addition, the members of the open data portal working group can give more detail on the functioning, procedures and use of the portal while the media sector representatives can give an overview of more media sector specific aspects. With those interviewees in mind, the saturation of information is aimed to be achieved, meaning that different sides of the questions should be covered by these different perspectives offering rich data.

The questions are divided in three sub-groups concerning each of the research questions. Based on the existing literature, the questions have been formed taking into account relevant aspects concerning the governmental view and the business view. For the governmental view, Toots et al (2017) and McBride et al (2018) have inspired the more general questions on benefits, barriers and drivers while for the legacy media business side, Susha et al (2015) have provided some skills and applications related questions. Perhaps the best framework to use for exploring these different aspects has been presented by Dawes et al (2016), namely their open data ecosystem model that highlights the importance of all stakeholders and the different sub-themes in view of open data provision and use, i.e. communication, the data itself, guiding policies, stakeholder interaction (19). Their ecosystem model is more broadly based within the sociotechnical studies theory (STS) which for short stands for a framework highlighting complex social, institutional and technological relations in view of a phenomenon (Dawes, Vidiasova and Parkhimovich 16). Here the complex phenomenon is the open data portal, its technical aspects but also the relations to and among stakeholders responsible for the maintenance of the portal and the stakeholders using data for their specific purposes. One thing the ecosystem approach lacks, is the barriers and drivers aspect that can be seen as an overarching layer that is relevant for all the elements that belong under the ecosystems model.

The questions of the first sub-group deal with benefits from open data. This serves as a more general topic, which in view of the interviews should help to establish that somewhat less detailed questions are asked in the beginning. Based on Trainor, starting out with more general questions is helpful for the interviewees and helps to establish rapport and reciprocity, which is deemed important in interview research aiming for rich inputs (Trainor 129-30). The second and third

(14)

sub-13

groups of questions are closely connected as topics- barriers and drivers to use open data. As Toots et al in their study suggest, what accounts as a barrier then in the opposite it very likely operates as a driver (105). Therefore, the relation between barriers and drivers is somewhat inversely proportional, meaning that having certain barriers means most likely not having certain drivers. As a result, it makes sense to consider this a single category of questions.

Under benefits, the question of private and public value can be addressed. Also, more specific detail on the application of open data in the media sector, impact on reader numbers, adverts, can indicate the financial benefits. The issue of public value can also be addressed by exploring what type of data has been used, to what purpose establishing whether this data is more used for stories on politics, and society or for entertainment purposes. These last questions are specific to the media sector while also other particpants’ views on the dual responsibilities of the media sector can give an indication on the expectations towards the media sector. Also, as Toots et al (2017) and Radchenko et al (2014) suggest, the lack of business models has been established problematic for the private sector, then exploration of potential open data uses in view of creating a business case can be beneficial. The second sub-group on the barriers and drivers of open data use concentrates on infrastructure, data, skills, regulation (privacy, security) issues exploring what exactly in each of those categories poses difficulties and opportunities. Annex 2

. Interview Questions

includes the full questionnaire.

With each issue in mind it will be possible to map which barriers and drivers impact more strongly the use of governmental open data by the media sector. While the media sector representatives will give a grass-roots practical account, the governmental and the non-governmental representatives’ input will serve as a reference point indicating whether the concerns match and whether there are media sector specific issues that in the future could be addressed by the policymakers and other stakeholders, i.e. the local media sector itself.

2.4. Ethics and limitations

The interviewees are anonymised but given that Estonia and the respective field is very small then despite the best efforts people familiar with the topic could identify them. This risk was presented to the interviewees. Consent for recording the interviews was asked to allow for later verification if needed. The interviews were stored on a personal hard drive to prevent access by unrelated parties and to maintain the confidentiality of the interviews.

There are some limitations to the methods. I have aimed to sample the documents and the interviews in a way that the received information reaches a point of saturation. This means that considering the scope of the study, the data sources are believed to offer thorough input. However, as is the case with qualitative studies, there is a question whether all useful data points have been covered and whether the analysis is reliable enough (Yin 96). The flexibility of semi-structured interview via reciprocal feedback with the interviewees as a method should diminish this risk as it allows for covering aspects that may accidentally be left out at first. Additionally, the ingrained beliefs of the researcher about whose voice matters (Trainor 131) and whether the chosen voices reflect the issue sufficiently is an issue. However, reflecting with experienced experts of the field about the suitability of the documents and the interviewees should reduce that risk.

Next, findings are presented. First, findings reflecting the governmental view on open data and the portal are presented and secondly, legacy media specific issues relating to open data are addressed. As said, the documents form an important framework while the interviews present more in-depth views concerning the media sector and its use of open data. As the documents depict the governmental view on open data more generally, the findings based on the documents help to form a more general view on the open data ecosystem and issues that are universal across sectors. The second findings chapter delves into the legacy media specific issues regarding benefits and barriers of open data.

(15)

14

3. Governmental Prioritisation of Open Data in Estonia

The first chapter on findings gives a more general picture of the open data ecosystem in Estonia. It studies two policy documents, the meeting minutes from the governmental open data working group and when relevant contrasts the findings with some of the statements from the expert interviews and the existing literature. For the purposes of answering the research question on how the media sector uses open data, it gives an overview regarding more general issues around open data and the open data portal that impact all sectors including the media sector. More specific answers to the question of potential benefits and the existing barriers for legacy media will be covered in the next chapter which is based mainly on expert insights.

3.1. Open Data Policy Developments

Mainly, two documents were studied. The first one is the Digital Agenda 20204_{from 2018 that is}

a more general policy document addressing broader development plans for the information society in Estonia. The second document is the Estonian Government’s Green Paper on Open Data that will be studied later. These two policy documents are the main strategic documents by the Estonian government frame the open data issue in the country. Therefore, the documents contain important information regarding the governmental perspective- to what extent is open data prioritised and the main issues concerning open data provision. Understanding the governmental perspective is important as it acts as a driving force behind the open data portal and with that in mind it will be a source to gather more general findings on the benefits, drivers and barriers of open data. Moreover, as highlighted in the literature review as per Mazzucato (2018), public sector has a significant role in investing into innovation and therefore its position regarding enthusiasm towards open data is important to study.

In the first document ‘open data’ is mentioned twice. It is used in relation to the third activity area- stimulating innovation under which open data is to be available in machine-readable format and promoted among the public for wider awareness (16). In the Estonian version of the document open data- avaandmed- in different cases is mentioned ten times. First, it is mentioned in the Trends chapter several times where the focus is on simplifying access to data, reuse of data to improve inter-institutional and international cooperation but also to encourage businesses to innovate based on open data. Comparing the coverage of the topic with other topics, open data does not seem to be a very prominent issue at the time of writing (2014). Considering the low scores in open data rankings at the time this finding is not surprising.

‘Linked data’ is introduced as a term and allows for faster incorporation of bigger amounts of data by applications which improves the level of analysis that can lead to better decision making (18). Secondly, it is mentioned under chapter 4.2 on developing the state’s information system where one of the directions is simplifying access to public information and offer solutions to use that data (26). On the next page, ‘machine-readability’, ‘awareness on open data’ and an ‘action plan for open data’ is suggested. What is important and was later confirmed by one of the interviews is that there is a principle that the state’s all new information system developments need to include a solution to provide open data as a default setting. Therefore, the principle of ‘open by default’ is being applied. A discrepancy arises with the mention of the open data related action plan as this is something that has not been realised by now, however several interviews confirm that though there has not been a concrete action plan since the Green Paper on Open Data from 2014, there will be one by the end of 2020. Further, the availability in relation to cultural data is mentioned (31) and finally, access to health related research data is suggested (42-43). Therefore, some open data related issues are raised and improvements are foreseen and were later continuously developed as confirmed by the interviews. However, problems were raised in the interviews concerning open data’s machine readability and low awareness. Therefore, there are some mentions but the topic is

(16)

15

not a top priority as issues relating to connectivity, ICT skills and smarter governance for example have full chapters written on them and though the links with open data provision can be imagined they are not explicit nor direct.

Green Paper on Open Data5_{was published by the Estonian Government in 2014. It is the first}

and up to now the last document that specifically deals with issues of open data on the governmental level. For the purpose of this paper, not all sections deserve full analysis. As the focus is on establishing the role of the media sector as one sector among private companies with the potential of using open data for innovation purposes, the section on impact and benefits of open data (section 3) deserves more attention, as well as the description on the state of affairs (section 7) and possible solutions (section 8) to improve open data policies. Section 3 starts with listing transparency and innovation as the first two main benefits with impact on activation of (data) economy as the third potential benefit followed by social impact, making the public sector more efficient and international obligations coming from the EU and keeping up the image of e-Estonia. The statement of open data as new gold is mentioned under the potential economic benefits, however the benefit of innovation seems to be linked to innovation in the public sector itself. Therefore, the benefits for the private sector are mentioned in the two points, however the elaboration is not very lengthy. Both private sector and also specifically the media is mentioned under potential benefactors (9). Media is mentioned in view of being able to access open data and influence the respective domain. It remains unclear whether dissemination of relevant data to the public is meant or if media is considered as a watchdog that is expected to influence decision making processes. From the wording, the latter seems more plausible. This relates to how open data is generally approached in the literature, meaning that it adds to transparency and accountability and that innovation as a topic serves a secondary purpose.

Under potential private sector benefits, integration of open data to create and improve services and applications is mentioned. It is stressed that open data provides a helpful addition to company’s own data. As per the interview with a data analytics expert from one media organisation this potential is seen but not used. However, this statement also relates to data journalism as an improved service. Of course as Gray et al (2012) suggest, a nose for good stories is also needed meaning that a journalist is the one who establishes the importance of an issue and then includes data to illustrate complex news and conveyed messages more strongly (Gray 2). Data and therefore open data can both be a source and a tool for a story (Gray 2) which broadens the point made in the Green Paper stating that open data is mainly for additional use.

Chapter 7 of the Green Paper describes the situation concerning open data in Estonia. It is worth mentioning that the description is from 2014 so the conclusions drawn from the document are very likely somewhat outdated. What the chapter does is present legal bases for managing open data related affairs. Topics of asking a fee, machine-readability clause, obligations for re-usability of data, the legal ground for the open data portal is described via references to the Estonian Public Information Act and the corresponding sections in the law. Further, issues of metadata, licensing, the requirement of ‘open by default’, open formats, interoperability are referred to. Also, other guiding documents like the Digital Agenda 2020 are mentioned (Ministry of Economic Affairs 2 23-6). Based on these issues, it seems the Green Paper was strongly guided by the Public Sector Information Directive (the PSI Directive) from 2012 where all of these issues were meant to be addressed by the EU Member States. The topics of the Green Paper that reflect the changes in the Public Information Act seemingly correspond to the changes made in the PSI Directive, mainly Articles 2, 3, 5, 7 and 9 that revolve also around charges, metadata, availability (openness), accessibility and such. This seems to show how the enthusiasm of the EU level and more directly the European Commission led the national developments in creating legislative and political

5

(17)

16

grounds for open data. The positive impact of a regulative framework was also mentioned by Toots et al (2017) confirming its importance for increasing open data use and with that potential benefits. Further, as discussed later on the basis of interviews the mentioned issues still pose problems, thus the Green Paper has some relevance even today as metadata, interoperability, data formats and machine-readability still require attention and improvements. It also confirms Barry & Bannister’s (2014) claims that data and data release barriers remain relevant (131-33).

The enthusiasm and pressure from the EU seems to drive the current processes around open data policies of Estonia also now. The newest PSI directive from 2019 is being implemented and from the minutes of the working group on open data it is visible that the developments are discussed from the PSI directive implementation perspective. More specifically, there are discussions around establishing what constitutes high value datasets that is being agreed among the Member States. Also, updates on other aspects of implementing the PSI directive are mentioned throughout the meeting minutes of the working group. This seems to suggest that there has been little time to concentrate on navigating the user perspective, including the media sector. The modest exploration of the user perspective by the government was also confirmed by the public sector representative, however, improvements in that area are foreseen and new developments will be discussed with experts from the private sector in relation to the new open data strategy that is compiled within 2020.

When it comes to examples of open data that can potentially be used in new applications and services, the Green Paper mentions geo-data, environmental, business and statistical data (Ministry of Economic Affairs 2 28-9). Some other burning issues that can be seen as challenges are that open data release is project based and the projects often have low priority (30). However, the priority has been increasing in recent years with developments concerning the national project KrattAI aiming for AI-based virtual assistants that make use of open data. This was raised by the OKE representative in the interview. Also, interoperability as a challenge is mentioned stating the need to set clearer standards to public organisations publishing open data echoing Barry & Bannister’s (2014) similar concerns (133). More broadly, the uneven level of provided data is discussed and the need for improvement stated (30). This was also one of the main issues as per Conradie & Choenni (2014).

One of the driving aspects is the rather active open data community like hackathons organised by an NGO called Garage48 that promotes inventing digital solutions to governmental, entrepreneurial or societal problems. Another issue is resources, mainly funding to increase the capabilities of the organisations to publish data that has good quality and that is relevant (31). The resulting needs stemming from funding are increasing the number of officials capable of providing solutions in publishing data and improving their skills (31). The last two issues are also mentioned in relation to public service co-creation via open data by Toots et al (2017: 104) confirming the relevance of these factors more broadly. Therefore, there are several problematic issues with one driver that the document lists- uneven level of data, low priority (which has improved), lack of funding, low skills versus an active open data community as a driver.

Additionally, the Green Paper lists eighteen more specific problems in view of open data. The full list can be seen in Annex 1. List of problems from the Green Paper on Open

Data

, however for purposes of clarity, the problems were divided into seven corresponding categories. The categorisations can be verified when looking at Annex 1. Mostly, it seems, the problems at the time were caused by lack of regulations and guiding documents to clarify procedures and rules (33-34), for example rules on interoperability, and lack of procedural rules on managing and developing the open data portal. Also, the organisational aspects within the public sector are mentioned, examples include low global and local interoperability, lack of setting clear priority areas, and dividing general organisational responsibilities of the stakeholders. Resources (finances) is another obstacle. Next come infrastructural issues such as lack of a central repository or a fully functioning portal that

(18)

17

aggregates and links to organisational datasets. Data specific issues cover low machine-readability and problematic formats of data. Modest awareness and use among both the providers and users of open data is listed. The user side is mentioned in view of their lack of experience by not participating in international projects and lack of new technological solutions based on open data. All of the problems have been covered in the literature and by the interviewees allowing for a conclusion that these issues need further improvement despite several accomplishments in the recent years. With that, the priority of the topic has increased due to more active communication and awareness raising, additionally rules have continuously been added and clarified to offer good quality data more broadly.

To compare the 2014 problems with what has been dealt with from then on and what is reflected in the working group minutes, it seems legal issues have been the highest priority- creating legal grounds, transposing the PSI directive, dealing with licenses have been on the agenda of the meetings and have also brought results. As according to Toots et al (2017) this facilitates use of open data the priority of having a clearer regulative framework serves the purpose. Other topics receive less attention, questions on data specific issues, infrastructure, use are also touched upon to an extent. Based on the meeting minutes, awareness is dealt in connection with data journalism and resources receive little attention. As the literature concerning the relation between mainstream media and open data relates strongly to data journalism this emphasis is not surprising.

Before turning to the current state of affairs it is useful to go over the solutions chapter (8) from 2014 as it allows to reflect what gaps have remained after the publication of the Green Paper, also in relation to the media sector. Chapter 8 states four main issues to be addressed. The first one is improving the organisational set-up by agreeing on concrete responsibilities via rules and regulations; operating and developing the open data portal as an organisation (a working group) but also as an infrastructure (links to databases); and monitoring the implementation of the strategy. Secondly, setting up an open data working group is stated and this has been in operation from 2018. These two tasks- organisational and infrastructural setup- have been delegated to Open Knowledge Estonia (OKE) signifying the positive developments regarding the two goals. Thirdly, organisations holding public sector information were expected to write an action plan for open data and this has been incorporated to the process of renewing information systems as ‘open by default’ is a requirement to develop new information system solutions as described by the interview focusing on the governmental perspective. Finally, financing is listed where it was foreseen that the administrative burden and costs in answering single public information requests will be balanced by a more comprehensive ‘open by default’ approach. This as a principle has according to the interviews also been taken up and thus reduces access barriers. Here the organisational questions have been since addressed, both the portal and the administrative side has been improved. Funding is a difficult aspect to assess as there have been contracts to outsource some of the administrative tasks and also calls for tenders concerning the open data portal have been posted. Therefore, though there can always be more funding it seems that the level of resources has improved since 2014. In short, all of the policy directions have been taken up to improve the open data ecosystem.

3.2. Open Data Policy Working Group Priorities

As the official policy documents offer limited insight into how the private sector and among that media is included in the development of a comprehensive open data ecosystem, it was necessary to study the open data working group meeting minutes6_{where governmental officials, a few}

scholars and Open Knowledge Estonia (OKE) participate. There were six publicly available documents describing the discussions in the working group. The meetings took place in 2018 and 2019, three meetings per year respectively.

6_{Meeting minutes can be found on}_{https://opendata.riik.ee/sundmused-ja-infomaterjalid/}_{and were last accessed}

Open Data Policy in Estonia: An Opportunity for Innovation in the Media Industry?