“Your data is only as good as what you do with it and how you manage it” [1]. This quote summarizes the problem that eGovernment Switzerland is currently facing.
Infrafederal Data Governance is needed
Many data are collected and stored in a more or less decentralized way from different public administrations at the three levels of the State (federal, cantonal and communal). The current system lacks efficiency, however, mainly due to the following reasons:
- The lack of centralized overview of existing data collection and variables already available for public administrations. This overview can be obtained by the creation of a metadata system that acts over time as a centralized catalogue of available data. The data protection level of each variable should also be indicated in this catalogue as well as the data ownership.
- The structure of the legal basis. The focus point of Swiss laws is to authorize the collection of data for business specific purposes like health, social, statistic, finance, transport and so on. This public policies oriented collection of data is a top-down approach which is entirely understandable and should stay mandatory (public administrations should only collect data needed for achieving their tasks if they do not already exist in official data collection). However, these public policies oriented and decentralized collection of data also creates data silos. Many data are collected in different contexts by different agents but are virtually identical as they concern the same data subject.
- Access for public servants to existing official data collections needs to be simplified as often as possible. Where personal data are concerned, data protection legislation needs of course to stay the first principle to manage access to these data but the creation and maintenance of data silos should not be possible without a very transparent and clear reason.
The inefficiency in the way how we collect and manage data today in order to achieve the tasks of public administrations does augment, not reduce the administrative burden of respondents (i.e. citizens, enterprises, authorities). Switzerland recently signed the European declaration on eGovernment (Tallinn Convention) [2] and its six principles show the right way. Data governance is a key issue that eGovernment Switzerland needs to face. With the revision process of the national eGovernment strategy launched, now seems to be the right moment to define better data governance in order to make the entire system more efficient.
What do we need to address?
The main reason that could explain the above mentioned problems is the absence of an existing National Data Governance (NDG). National data governance refers to governance which applies to the three levels of public administrations (federal, cantonal and communal). Public administrations need to address this challenge by implementing centralized data governance and at least a centralized data and metadata architecture but decentralized data management driven by business and not by IT.
This kind of governance does not exist at present and should now be officialy implemented as soon as possible. The following first steps could be adressed:
- The implementation of a National Data Quality Management (NDQM) that should help public administrations to structure their activities around Data Governance. Frameworks, competence centers (e.g., competence center Corporate Data Quality) have the maturity to help public administrations to implement this NDQM. Topics such as Master Data Management (MDM), Data and Metadata Architecture, should be adressed under this umbrella.
- The implementation of a National Data Infrastructure which could be defined as “a network of linked official administrative registers built over base registers (People, Enterprises and Location) representing the core vocabularies and managed under a comprehensive classification scheme; accessed and updated on the basis of authorised administrative roles and acquired personal roles for an integrated, version controlled, auditable, enterprise wide records management system.” [4]. Standards for data exchange between this base and administrative registers already exist, e.g., eCH and unique identifiers (UID, EGID, EWID,…). Without this National Data Infrastructure, it seems just impossible to embrace the challenge of reducing the administrative burden of respondents. Only the variables that are not already existing in this infrastructure should be collected. The Open Government Data Platform should also be seen as a building block of this National Data Infrastructure. The use of Cloud and Blockchain technologies need to be integrated into this system and not be treated as an issue apart.
Are we ready for Data Analytics?
Big Data (social network, satellite imagery, …) and the Internet of Things (IoT) are current buzz words, but they are also referring to a new ecosystem that has dramatically changed over the last five years. The increasing amount of available data that governments could and should have access in order to support decision making is growing exponentially [3]. These data are often in the hand of private companies that are more and more located neither in Switzerland nor Europe. These private companies do not need to initiate surveys to collect all this information, it is part of their digital value chain and business model. How should public administration react to this new reality? Can civil servants embrace these new challenges in the light of an established National Data Governance and a National Data Infrastructure?
Today different legal bases (include data protection laws) exist depending in what context the data are generated. The skills, knowledge, and infrastructure that public servants will need to collect, integrate and analyze all thisstructured and unstructured data (Analytics) do not figure on the roadmap of the eGovernment Strategy 2016 – 2019. There is a certain need to rethink these priorities.
Analytics in the context of Big Data is not just an extension of existing Business Intelligence concepts but a fundamental change regarding the algorithms that can and should be used (advanced statistics, data science, machine learning, deep learning and so on).
The skills and the knowledge that public servant will need to collect, integrate and analyse all those structured, semi-structured and unstructured data is more or less not existant in the public administration. It needs to be developped quickly. Governments at the three levels of federalism need to develop the needed hard and soft skills [5] to work with data in order to achieve their tasks. The accurate capacity to analyse the “right” data quickly in a correct and transparent way is key in this context.
Should Data Governance be part of the eGovernment Strategy?
The answer seems to be clearly yes. So let us stop dabbling and make use of the current revision process of the Swiss national eGovernment Strategy to tap the potential of data for our society and economy, and let us treat data as a key strategic asset [6], so ensuring their veracity the related data quality and data governance become imperative.
[1] Judith Hurwitz, Daniel Kirsch, Machine Learning, IBM Limited Edition, 2018, pp. 1.
[2] European declaration on eGovernment, Federal Department of Finance, 2017. https://www.efd.admin.ch/efd/en/home/dokumentation/nsb-news_list.msg-id-68342.html
[3] Banning Garrett, Big Data is changing your world, ETHZ, 2013. http://www.css.ethz.ch/en/services/digital-library/articles/article.html/173004
[4] Data-Driven Public Administration, National Data Strategy – Malta, 2016, pp. 2. https://mita.gov.mt/en/nationaldatastrategy/Documents/Data-Driven%20Public%20Administration%20(Malta).pdf
[5] Open Data Maturity 2016, European Commission, Directorate General for Communications Networks, Content and Technology, 2016, pp. 65. https://www.europeandataportal.eu/sites/default/files/edp_landscaping_insight_report_n2_2016.pdf
[6] Enter the Data Economy, EU Policies for a Thriving Data Ecosystem, European Commission, European Political Strategy Center, 2017, pp. 1. https://ec.europa.eu/epsc/sites/epsc/files/strategic_note_issue_21.pdf