Chapter 6. 6 Open Government

Open Government, Open Data

Moore’s law states that the power of computer chips doubles every 2 years, and this also applies to computer data volumes. Faced with these challenges, governments are turning to cloud computing, taking the user to where the data are found rather than the other way round.2

At the same time, several public administrations are joining the open government movement3 (Estermann, 2018). In 2011, the Government of Canada launched its first National Action Plan on Open Government. It was one of the first countries to join the Open Government Partnership (OGP), a multilateral partnership endorsed in 2011. As of June 2024, 75 countries and 150 local governments—representing more than two billion people—as well as thousands of civil society organizations are members of the OGP. This is a global movement.

The first area developed during the implementation of open government is access to data, commonly known as open data.4

A Multitude of Data Portals in Public Administrations

Open data from public administrations is usually grouped together and made available on a single portal. These data are listed in a catalogue that can be searched using website-specific nomenclature and search tools (Tygel et al., 2016). The Government of Canada presents its catalogue of open data on its website, where approximately 40,000 metadata available can be searched.

Each government portal offers tools for identifying and sometimes visualizing datasets. These are often grouped by theme, such as health, culture, or technology (Tygel et al., 2016).

Government portals sometimes bring together descriptions of datasets from several administrations.5 In Canada alone, there are at least 66 different open data portals.

There is a lack of uniformity in the interface, functionalities, and categories across these portals, many of which are only available in English, which limits both discoverability and access to data for people who do not speak this language.

To find a dataset specific to each government, users must therefore discover the portals, familiarize themselves with the search tools, and identify the dataset in question, and only if they are fluent in English. The opening up of data is a step in the right direction for open government, yet it has to be said that the existence of a multitude of portals creates silos restricting access to data.

Centralizing Metadata on Wikidata

The Wikimedia Foundation was created in 2003, 2 years after the birth of Wikipedia. Its aim was to fund technical support for the free encyclopedia. Gradually, several projects were added, all with the aim of making knowledge accessible to everyone (Yoakim, 2020).

Wikidata was launched in 2012 by the German chapter of the Wikimedia Foundation. It is a free, multilingual, collaboratively edited knowledge base that provides answers to complex queries and is part of a context of linked data, the basis of the semantic web. Its content is licensed under the Creative Commons Zero (CC0) licence,6 which encourages the free reuse of the knowledge and information it contains.

By June 2024, more than 10 years after its creation, Wikidata had more than 110 million items and more than 23,000 active users. Its data can be viewed using 50 external tools.7 More than 500 articles have been published about Wikidata.

Several open datasets (including Wikipedia’s) have been integrated into Wikidata, in all languages. Wikidata is conducive to the centralization of government open data, not least because of its semantic web capabilities (Ajose-Ismail & Osanyin, 2019).

These capabilities are all the more interesting given that Wikidata contains information, for example, on administrative regions (1.9 million entries), chemical compounds (1.2 million), and communication channels (630,000 entries). Metadata on scientific publications account for over 30% of entries, with 22.6 million descriptive records (“Wikidata: Statistics,” 2025).

Wikidata makes it possible to add structured data using scripts, and it offers a protocol (SPARQL) for searching for, adding, modifying, or deleting data. Wikidata constitutes a solid infrastructure capable of absorbing millions of items of metadata, as the example of scientific articles shows.

Example of the Meteorological Service of Canada

Historically, the exchange of meteorological data has been an area where collaboration between countries and scientists has always predominated. Indeed, a weather forecast over more than a 48-hour period has to integrate data from all over the world. And as this is true everywhere on the planet, it encourages countries to exchange data since they all need data to make forecasts for their own territory. Even at the height of the Cold War, meteorological data circulated between Eastern Bloc and Western Bloc countries (Zillman, 2019). This is one of the reasons why this area of knowledge is particularly well suited to the Wikimedia Foundation’s ideal of knowledge sharing.

In 2019, the Meteorological Service of Canada granted funding to Wikimedia Canada to upload weather station observation data and metadata to the Wikimedia Foundation ecosystem.8 These data are uploaded to the Wikimedia Commons and Wikidata, respectively (see figure 6.1). The metadata includes a link to the corresponding weather station data in the commons. The metadata for Canada’s weather stations are added to those for 12 other countries with more than 10 stations in Wikidata.

It is now possible to use the tools and exploit Wikidata’s various calculation and query capabilities. For example, you can cross-reference station metadata and then create categories of stations by altitude. You can also cross-reference the metadata with other information already available in Wikidata and, for example, identify weather stations in Québec located within a radius of one kilometre of a bridge longer than 100 metres. The information on bridges already existed in Wikidata when the station metadata were imported. These few examples illustrate the power of Wikidata’s search tools.

In connection with this project, Wikimedia Canada and Acfas welcomed two Institut de valorisation des données (IVADO) interns in the summer of 2021 from the Data Storytelling program.

Two articles were based on the collaborative work of Laurence Taschereau (UQAM, journalism) and Ali Akbar Sabzi Dizajyekan (Polytechnique Montréal, data visualization) on meteorological data. Their work is helping raise public awareness of the effects of climate change on the population. The first article deals with agricultural resilience (2021b) and the second with heat islands (2021a).

Wikimedia Canada and the World Meteorological Organization (WMO) are working together to find a way of uploading all the metadata from weather stations registered with the WMO to Wikidata. With a view to creating weather analysis products, exchanges are also taking place with users in Cameroon and Nigeria, countries where IT infrastructure capacity is less developed.

First Stage in Importing Data into Wikidata: Targeting Metadata and Standardization

Each open government follows a certain standardization to describe its datasets. The information often includes but is not limited to the licence, the administrative entity (government ministry) responsible for the dataset, the URL for downloading the data, and the date of the last update. The Open Data Charter is working to establish an international charter for open data, including interoperability, so that data can be compared and cross-referenced, regardless of the original source.

The aim is to convert catalogue metadata, which are standardized locally, into a form that corresponds to that metadata stored within the Wikidata database.9 Some metadata lend themselves better than others to being imported into Wikidata, as they adopt an international standard as well as a local one. Converting them to a common format in Wikidata is therefore less complex.

Any international standard is supported by a community made up of users and contributors who are likely to take part in the project to adapt the catalogue for import into Wikidata. In addition to this meteorological data, which meet this criterion, georeferenced data serve as another example of standardized data. These metadata describe administrative areas (political and census) as well as routes for different means of transport, among others.

Taken together, these metadata could make it easier to deploy a host of applications on a global scale, which is currently impossible given the need for time-consuming manual processing of metadata.

Benefits of Centralizing Open Data Catalogues

Open governments have centralized the data of their ministries and agencies under a single portal, imposing a standardization of dataset metadata, as explained earlier.10 The centralization of Wikidata11 is a simple extension of this principle: creating a single place where not just one but all the catalogues of open governments can be found, making them accessible to citizens all over the world.

Cloud Infrastructure and Community Already in Place

Wikidata is a database available on a reliable infrastructure, financed and managed by the Wikimedia Foundation. The platform is regularly monitored, and capacities are increased as required, which includes integrating new technologies. Governments wishing to contribute their metadata will benefit from existing servers, and users will have free access to the data.

Although it is difficult to estimate how much Wikidata would contain if governments added their metadata—consider, for example, that the Government of Canada has 40,000 entries in its open data catalogue—it is reasonable to think that the infrastructure would be able to tolerate these massive imports. By way of comparison, the number of entries in Wikidata for scientific articles is over 22 million. Following a massive import of elements, it is vitally important to maintain dialogue with the Wikidata community to avoid any spontaneous, sharp reaction. The incremental nature of the import process will leave room for negotiation in order to satisfy the community.

The results of user queries in Wikidata can be displayed in a number of ways: in tables, graphs, or maps. The results can be exported in a variety of formats, and the data can be reused in other systems.

Finally, there are a number of external applications for carrying out searches or calculations in Wikidata. A directory page lists these applications, which often work by thematic field (“Wikidata: Tools/Visualize Data,” 2026). It is possible to develop specific tools for open government metadata.

The Challenges of Centralization

The first and potentially most important step in copying metadata into Wikidata is to create a correspondence between the typology of open data catalogues and their representation in Wikidata. This requires knowledge and expertise in the content of both the catalogues and Wikidata. Such a task requires human resources and time, and it can extend to the creation of properties in Wikidata in collaboration with the community.

Once the matches have been created, the metadata is uploaded to Wikidata, ideally using automated scripts. This task requires programming skills, which are key when it comes to connecting a catalogue to Wikidata.

When the catalogue is updated on the government portal side, the information needs to be synchronized in Wikidata, which means uploading the metadata that have been added or modified. This can be done at regular intervals (every week, month, or year) or as and when required. Conversely, a feedback mechanism should be put in place to communicate the changes made in Wikidata to those responsible for government metadata. The community can detect errors or incongruencies in the metadata once they are being used (Krabina, 2023). The element tracking system, already present in Wikidata and mentioned earlier, can be used to provide this feedback.

The key to the massive import of government data into Wikidata is therefore regular communication between the community and its government counterparts. The ideal is to obtain the support of a community of practice with which to establish links of trust. The source code of the scripts should be free to share best practices.

Granularity

When it comes to managing metadata, the question of granularity is omnipresent. At what level should datasets be represented? By grouping them together using the lowest possible level of abstraction or, on the contrary, by going to the highest possible level of granularity? Both solutions are equally valid.

To take the example of weather stations again, should we publish a single metadata entry containing all the 8,000 or so stations listed, or should we publish one entry per station? Or both? The difference in the volume of metadata to be entered in this case is several orders of magnitude.

One of the advantages of the higher level of metadata granularity is that Wikidata’s tools can be used to their full potential to identify data. This requires extensive work to establish the metadata correspondence between the government catalogue and Wikidata.

It would be surprising if a general rule could be identified. Metadata owners must make choices. As with categorization, this could be undertaken in consultation with the fields of interest community, where such a community exists.

Conclusion

Initial interest in the meteorological metadata project was surprising. We were contacted by governmental and paragovernmental organizations as well as civil society institutions interested in open knowledge (libraries, NPOs, private companies). This growing interest has encouraged us to extend our thinking to other datasets, since the principles underlying the massive import and dissemination of meteorological data can be applied to them as well.

The notoriety of Wikipedia and other Wikimedia Foundation projects promotes understanding and acceptance by government officials. The attractiveness of using existing IT resources is an asset, where their methods of access, search, and dissemination are recognized by public administrations and open data users in general.

An infrastructure of this kind also facilitates collaboration between different levels of government or even between states because the issues of infrastructure costs and, above all, the policy of managing and funding this infrastructure are in the hands of a third party: the Wikimedia Foundation. All that remains for governments to do is invest in the human capital needed to use this infrastructure and, of course, to answer the ontological questions that cannot be ignored.

We have sketched out the advantages and challenges of centralizing metadata in Wikidata. The next step will be to identify the metadata whose sharing could benefit society as a whole. We will then need to invest in the longer term to explore and mark out the path leading to greater dissemination of open data.

Discoverability issues are the main obstacles to the dissemination of open data. Investing in new solutions and joining a community or existing projects are actions in line with the values underpinning the open democracy movement. All that remains is to take the second step.

Generative AI tools, such as ChatGPT, will help mine the data to answer various questions. However, for these tools to be truly effective, it is essential that the basic information be accessible. It is crucial that these data are reliable and sourced to ensure that the answers provided are based not solely on statistical inference but on accurate and verified data. So accessibility and quality of data are essential prerequisites for taking full advantage of the capabilities of generative AI in the management and analysis of open data.

References

1. Ajose-Ismail, B. M., & Osanyin, Q. A. (2019). A semantic approach for facilitating search for discovery of open government data on open data portals. In Book of proceedings of the 4th National Development Conference of the School of Pure and Applied Science, The Federal Polytechnic Ilaro, Ogun State, December 2–5, 2019 (pp. 249–56). https://eprints.federalpolyilaro.edu.ng/1645/
2. Casemajor, N., Béland, J.-P., & Rochon, B. (2017). Comment dit-on page d’accueil? Incorporer le savoir autochtone dans Wikipedia. Diff. https://diff.wikimedia.org/fr/2017/05/15/creer-dune-encyclopedie-atikamekw-a-partir-de-zero/
3. Estermann, B. (2018). Development paths towards open government—an empirical analysis among heritage institutions. Government Information Quarterly, 35(4), 599–612. https://doi.org/10.1016/j.giq.2018.10.005
4. Evans, J. (2024). Aligning and sharing data with Wikidata—a feasibility study. Knowledge Commons. https://works.hcommons.org/records/5ym2e-bap78
5. Farda-Sarbas, M., & Müller-Birn, C. (2019). Wikidata from a research perspective—a systematic mapping study of Wikidata. arXiv. https://doi.org/10.48550/arXiv.1908.11153
6. Help: Sources. (2026, January 19). In Wikidata. https://www.wikidata.org/w/index.php?title=Help:Sources&oldid=2456148705
7. Hitz-Gamper, B. S., Neumann, O., & Stürmer, M. (2019). Balancing control, usability and visibility of linked open government data to create public value. International Journal of Public Sector Management, 32(5), 451–66. https://doi.org/10.1108/IJPSM-02-2018-0062
8. Krabina, B. (2023). Building a knowledge graph for the history of Vienna with semantic MediaWiki. Journal of Web Semantics, 76. https://doi.org/10.1016/j.websem.2022.100771
9. Krabina, B., & Polleres, A. (2021). Seeding Wikidata with municipal spending data. ResearchGate. https://www.researchgate.net/publication/355205810_Seeding_Wikidata_with_Municipal_Spending_Data
10. Lubiana, T. (2023). SPARQL queries, GPTs and large language models—where are we currently? Wisecube. https://www.wisecube.ai/blog/sparql-queries-gpts-and-large-language-models-where-are-we-currently/
11. Mora-Cantallops, M., Sánchez-Alonso, S., & García-Barriocanal, E. (2019). A systematic literature review on Wikidata. Data Technologies and Applications, 53(3), 250–68. https://doi.org/10.1108/DTA-12-2018-0110
12. Office québécois de la langue française. (2013). Données ouvertes. Vitrine linguistique. https://vitrinelinguistique.oqlf.gouv.qc.ca/fiche-gdt/fiche/26519745/donnees-ouvertes
13. Pellissier Tanon, T., & Kaffee, L.-A. (2018). Property label stability in Wikidata: Evolution and convergence of schemas in collaborative knowledge bases. In Companion of the Web Conference 2018 on the Web Conference 2018 (WWW ’18) (pp. 1801–3). ACM Press. https://doi.org/10.1145/3184558.3191643
14. Pellissier Tanon, T., Vrandečić, D., Schaffert, S., Steiner, T., & Pintscher, L. (2016). From Freebase to Wikidata: The great migration. In Proceedings of the 25th International Conference on World Wide Web (pp. 1419–28). International World Wide Web Conferences Steering Committee. https://doi.org/10.1145/2872427.2874809
15. Taschereau, L., & Sabzi Dizajyekan, A. A. (2021a). Des inégalités climatiques à rendre insomniaque. Acfas Magazine. https://www.acfas.ca/publications/magazine/2021/11/inegalites-climatiques-rendre-insomniaque
16. Taschereau, L., & Sabzi Dizajyekan, A. A. (2021b). La résilience agricole se trouve à nos pieds . . . dans le sol. Acfas Magazine. https://www.acfas.ca/publications/magazine/2021/11/resilience-agricole-se-trouve-nos-pieds-sol
17. Tygel, A., Auer, S., Debattista, J., Orlandi, F., & Machado Campos, M. L. (2016). Towards cleaning up open data portals: A metadata reconciliation approach. In 2016 IEEE tenth international conference on semantic computing (ICSC) (pp. 71–78). IEEE. https://doi.org/10.1109/ICSC.2016.54
18. Wikidata: Statistics. (2025, August 31). In Wikidata. https://www.wikidata.org/w/index.php?title=Wikidata:Statistics&oldid=2398684746
19. Wikidata: Tools/visualize data. (2026, January 19). In Wikidata. https://www.wikidata.org/w/index.php?title=Wikidata:Tools/Visualize_data&oldid=2455976888
20. Wikidata: WikiProject performing arts/fr. (2023, December 3). In Wikidata. https://www.wikidata.org/w/index.php?title=Wikidata:WikiProject_Performing_arts/fr&oldid=2023250515
21. Yoakim, W. (2020). Wikipédia, Wikimedia Commons et Wikisource, un Eldorado de visibilité. Archives, 48(2), 41–81. https://doi.org/10.7202/1067524ar
22. Zillman, J. W. (2019, November 27). Origin, impact and aftermath of WMO resolution 40. World Meteorological Organization. https://wmo.int/media/magazine-article/origin-impact-and-aftermath-of-wmo-resolution-40

1. 1 One of the special features of Wikidata is that it is based on a volunteer community. It was this community strength that led Google to migrate Freebase’s data to Wikidata. See Pellissier et al. (2016).

2. 2 See, for example, the DIAS platforms, funded by the European Commission, and the Australian government’s project Digital Earth Australia, which offers cloud computing space for the manipulation and posting of satellite data.

3. 3 There are several definitions of open government. According to the Canadian government, open government is defined as “a governing culture that holds that the public has the right to access the documents and proceedings of government to allow for greater openness, accountability, and engagement” (Government of Canada, 2014).

4. 4 Open data are “raw, non-nominative, royalty-free data produced or collected by a public or private organization, accessible to citizens via the Internet” (Office québécois de la langue française, 2013).

5. 5 For example, the directory of the Government of Canada includes data from the Government of Alberta. The directory of the Government of Québec includes data from municipal administrations in Québec.

6. 6 The CC0 license is a free Creative Commons license that allows copyright holders to waive as much of their rights as possible, within the limits of applicable law, in order to place their work as close as possible to the public domain.

7. 7 The list of tools can be accessed on the following page: “Wikidata: Tools/Visualize Data.”

8. 8 A test was also undertaken using financial data from Austrian municipalities (Krabina & Polleres, 2021).

9. 9 Some open software applications have been designed to host open data catalogues (Socrata, CKAN). More than 250 organizations publish their data using these open software applications using metadata standards such as DCAT or standards published by the W3C Government Linked Data (GLD) task force (Neumaier et al., 2016; Tygel et al., 2016). However, none of these data catalogues are retranscribed directly into Wikidata.

10. 10 Hitz-Gamper et al. (2019) have also drawn up a list of advantages and disadvantages but in a more general framework, focusing instead on government-linked data in the broadest sense. Here, we focus on the specific case of Wikidata.

11. 11 Although the semantic web was designed to operate in a decentralized fashion, there remain challenges of nomenclature among databases (Ajose-Ismail & Osanyin, 2019), and harmonization could be undertaken in Wikidata.

12. 12 A single point of failure (SPOF) is a point in a computer system on which the entire system depends and whose failure brings about the complete shutdown of the system.

13. 13 In Canada, government productions are subject to Crown copyright and could be subject to a special copyright regime. For more information on the subject, see https://ised-isde.canada.ca/site/ised/en/terms-and-conditions/about-copyright.