- June 2018: See m:Grants:Project/ScienceSource/Profile for the sequel project. Contact User:Charles Matthews for more. See also the related Wikidata:ScienceSource focus list.
WikiFactMine is a ContentMine project to add referenced scientific facts to Wikidata. It carries out searches of the scientific literature, using search terms divided up into sections called dictionaries. The approach taken is modular and allows participation.
WikiFactMine develops open software to bridge the gap between the recent scientific literature, held online in a repository, and database entries. The process is not completely automated, the final step being taken relying on human judgement. The project is pro-actively bringing in volunteers to work with the ContentMine system. We aim to find fruitful areas for collaboration, and to offer a daily update of facts.
This is the homepage for WikiFactMine on Wikidata. For a quick introduction to WikiFactMine, see the video from the WikiCite conference 2017.
A short welcome
The project is funded by the Wikimedia Foundation, and is intended to involve volunteers. Your dictionary can be here! WikiFactMine will work with you to help prepare and upload a dictionary of search terms, that will then be run daily over one or two thousand recent scientific papers.
WikiFactMine is in essence a large search engine, able to carry out multiple searches in parallel. All it takes to broaden the scope of the search is to prepare and upload a single data file, in JSON format. WikiFactMine will assist anyone interested in producing a dictionary. Existing Wikimedia content, on Wikidata or within the category system of the Wikipedias in various languages, is a fruitful source for dictionaries. Importantly, Wikidata items required to identify what is found in the search can be carried along.
Keep in touch
For support with and advice about the WikiFactMine and ContentMine software, contact User:T Arrow, for example at tomcontentmine.org. To hear more about our view in the large, sign up at w:Wikipedia:Facto Post mailing list.