For the participating libraries, the goal of the hackday series was to explore ways in which they could engage with Wikidata - both by contributing data, and by using it to enhance their own service offerings. Participants were expected to contribute around 150 hours each over a six-months period. In small groups of up to 5 people they worked on four different projects, three out of which were successful and are described in more detail below. A fourth one, aiming at mapping a correspondence network on Wikidata, was abandoned due to insufficient data.
Graph-based Query Expansion on Library Discovery System
A first project focused on enhancing the search experience on library discovery systems by drawing on Wikidata and the GND.
The ETH Library uses ExLibris Primo as its discovery system search portal through which users can search and access the library's resources and materials. The common way to carry out a search is to enter a search term, upon which a list of search results is returned. The result list can then be further filtered by facets. This approach has however a major shortcoming: It is based on a search for strings and not on a search for the actual things. Thus, the system does not allow searches for entries on specific places or persons, leaving it to the user to skim through the search results in order to separate the useful from the useless. At the same time, the demand from the Digital Humanities for more precise search possibilities has been increasing over the past years, and investments in respective digital infrastructures have been made.
This is where Wikidata comes in handy, as its graph allows more diverse information combinations than the discovery system. The Wikidata Hackday series was therefore used to explore ways to expand the Primo discovery system by drawing on Wikidata. From a technical point of view, Primo offers the possibility to add functions via AngularJS programming. The code of the prototype has been made available at Gitlab. A Documentation is also available there.
Improving the research possibilities in Primo by allowing for searches for precise persons and places leads to two advantages:
Users are provided with additional research possibilities, allowing to filter or expand search results around specific persons or places. – So far, the new functionality has been deployed on a test system; deployment on the productive system is expected over the coming months.
Referencing Archival Fonds on Wikidata
A second project focused on referencing archival fonds on Wikidata, so that the respective information would be available as linked open data and could easily be retrieved by third party services, such as library discovery systems or archival platforms offering searches across organizations.
Referencing archival fonds regarding persons on Wikidata is straightforward: it suffices to interlink the Wikidata item of a person with the Wikidata item of the respective heritage institution by means of the property "archives at" (P485). A qualifier "inventory number" (P217) is used to indicate the inventory number, and a "reference URL" (P854) to provide a link to the respective entry in the online catalogue (see e.g. Leopold Ružička (Q122996) for an illustration).
The participating institutions each ingested several hundred references to archival fonds about persons. In many cases, new person items had to be created based on the data available at the institution. This was done for all the persons for which the first name, surname, year of birth and year of death (if applicable) was known. In the case of incomplete data, the data ingest was postponed until further research would yield the necessary information.
The project team initially also looked into referencing archival fonds about organizations, but quickly realized that it would run into difficulties related to the heterogeneity of approaches when it comes to representing organizational changes over time in the source systems (name changes, mergers, acquisitions, etc.). This is an issue that needs to be tackled in the context of another project.
The project team is delighted to see that some Wikipedias already automatically transclude the information about the archival holdings in their person infoboxes (e.g. French or Catalan Wikipedia), giving the newly added data immediate visibility. The infobox templates could now be extended to also provide the direct links to the respective entries in the archival catalogues. Further outreach efforts are needed to convince other Wikipedia communities to automatically add pointers to relevant archival holdings from their Wikipedia articles (through infobox templates or by other means).
If you are interested in referencing your archival fonds on Wikidata as well, please find our step-by-step documentation on our procedure here: How to Link Your Institution’s Collections to Wikidata?https://doi.org/10.3929/ethz-b-000393724
Sum of All Swiss GLAMs
The third project tackled the implementation of the Sum of All GLAMs project in Switzerland. The goal of this project is to add data about all heritage institutions on Wikidata and to make this data available on Wikipedia through infoboxes. Keeping the data for all Wikipedias in one central location on Wikidata will ease the burden related to data maintenance in the future (both for Wikipedia communities and data providers). Including the data in Wikipedia infoboxes will enhance the data's visibility and thus draw people's attention to possible errors and incomplete data.
The hackday team's activities comprised the following:
evaluating the completeness of the data for Switzerland;
tackling data modelling issues and cleaning up data on Wikidata;
evaluating the possibility of ingesting data about library branches from the Swiss National Library's database of heritage institutions (IS Plus); and
engaging with colleagues at the Wiki Movement Brazil and various Wikipedia communities in view of the creation of infobox templates transcluding data from Wikidata (see example).
By doing so, it prepared the terrain in view of a pilot project to be launched in 2020, by means of which we are planning to invite Swiss heritage institutions to check, correct and complement their data entries on Wikidata. The case report of the Sum of all Swiss GLAMs project is available online.
In 2020, the three projects will be further pursued and promoted among Swiss heritage institutions and at various conferences. We found that the three approaches are complementary and reinforce each other and will promote them as part of a larger linked open data strategy. We would be happy to see them propagate to other countries and are looking forward to hearing from you if we have awakened your interest.