From Librarybase
Jump to: navigation, search

Librarybase is a project to develop structured bibliographic data around the citations that appear on Wikipedia. This data serves many useful purposes, including:

  • Facilitating research on where Wikipedia gets its information. For example, medical articles can be audited for their use of non-reliable sources.
  • Powering source discovery tools that help guide the work of Wikipedia editors.
  • Serving as a standard database for all the sources used on Wikimedia projects, including those that do not have standard identifiers such as DOIs.

How it works[edit]

The service runs on Wikibase, the same software that runs Wikidata. Each record created is that of a source, an author, or a type of source (journal, article, etc). Example records:

  • Q2 – a book
  • Q6 – a journal article
  • Q7 – an author
  • Q13 – a website

A script is under development to mine citation data from Wikipedia and import it en masse into Librarybase. Librarybase will likely start with a small subset of articles as a proof-of-concept before expanding into the rest of the encyclopedia.

Why not Wikidata?[edit]

Putting this data into Wikidata would be an excellent idea. Before this happens, it is important to make sure that:

  • The mechanisms for mining citation data and mass-importing it work fine;
  • The data surfaced is interesting and worthwhile;
  • Wikidata actually wants the data.

If all works according to plan and the data is worth incorporating into Wikidata, then plans for a migration will come together. If Wikidata rejects the Librarybase content, including the level of detail sought by Librarybase, then Librarybase will continue as an independent project.