Wikidata:WikiProject LD4 Wikidata Affinity Group/Affinity Group Calls/Meeting Notes/2020-11-03
Jump to navigation
Jump to search
Call Details[edit]
- Date: 2020-11-04
- Topic: Author Items and the Author Disambiguation Tool
- Speakers: Daniel Mietchen and Arthur Smith
Presentation Materials[edit]
- Call Agenda
- Wikidata:WikiCite
- https://github.com/arthurpsmith/author-disambiguator
- Author Disambiguator Tool
- Author Disambiguator Tool Wikidata Documentation
- Presentation Slides
Meeting Notes[edit]
- Notes:
- Author Disambiguator Tool creates a process for linking works to their authors in Wikidata, trying to semi-automate a process.
- Review of a scholarly article entry in Wikidata: some authors have linked author items with qualifiers series ordinal, some authors listed only as author name strings. “Affiliation string” qualifier is available as well. DOI most useful identifier for scholarly articles.
- 36 million scholarly articles, averaging 4 authors/work, 3.5 as a string, only 0.5 as a Wikidata item
- String to item conversion: link names unambiguously to individuals, account for name variants, be able to add useful information about that person to Wikidata, be able to use the query service to find useful information
- How to work with author items: manually, LargeDataSetBot (Orcid + bots), SourceMD. And the Author Disambiguator!
- Searching author names:
- Default search parses name into FN, LN, explores initial options for that name. Generates exact strings to help the SPARQL query process.
- Use Specify Name Strings to edit the list of names generated
- Filter ideas: journal, topics, etc.
- Results of name searches display in Groups, sorted articles based on co-author names primarily, then journals,topics. Also provides potential author items below groups.
- Searching by Works: will list all the authors in series ordinal order, links to author disambiguator page for each name, can edit some fields on the work item
- Can sign in with OAuth so you can edit items with your Wikidata login (vs. working through QuickStatements)
- Daniel: link out to Scholia from an Author view of Author Disam tool. Click to Scholia. Scholia has a missing page that can return you to the Author Disamb Tool. “Missing page” in Scholia to facilitate curation by author, topic, works, journals.
- Improving co-authors and topic matching will further improve clustering, so returning to topics, authors is a good strategy. Use one tool to improve the other!
- Questions:
- Via email: “any tips or tricks for working with authors with common names in author disambiguator. For example, I have an author Elizabeth Matisoo-Smith (Q16403266), for whom the tool insists on presenting every publication for an Elizabeth Smith first (none of which are the right ones). I cannot get it to use the Matisoo part of her name even though that would make the job a lot easier! And there is no way I can see to tell it 'none of these are right, give me the next 50 publications'! Anyway, I wondered if you might pass the question on in case Arthur or Daniel might have some tips that would make this sort of task easier.”
- Zoom in on just one of the groups provided in the Author Disambiguator
- Apply filters by journal or topics. Helps improve the clustering for future work
- Working on non-”article” formats -- works as long as the property “author name string” is present
- Via email: “any tips or tricks for working with authors with common names in author disambiguator. For example, I have an author Elizabeth Matisoo-Smith (Q16403266), for whom the tool insists on presenting every publication for an Elizabeth Smith first (none of which are the right ones). I cannot get it to use the Matisoo part of her name even though that would make the job a lot easier! And there is no way I can see to tell it 'none of these are right, give me the next 50 publications'! Anyway, I wondered if you might pass the question on in case Arthur or Daniel might have some tips that would make this sort of task easier.”