samwalton9 / wikicitevis Goto Github PK

We could allow searching for journal names (as strings), mapping to their DOI and returning results for it, in the same way Cite-o-meter does. We could even borrow their doi_pref file: https://github.com/wpoa/Wikipedia-Cite-o-Meter/blob/master/doi_pref.tsv

This could then be used to investigate the topic distribution of results.

We hit a row with the header values instead of actual data. Suspect we might accidentally be adding the header part-way through the file.

Currently the only options are 'All' or a single language. There might be use cases for searching a subset of languages.

Users would likely appreciate a 'Download as CSV/JSON' button next to the search results.

The citations data doesn't know how many times a citation was used within the article (i.e. with ) so grouping by article isn't exactly possible.

Above or below the search results, the 'count' should be shown.

https://unpaywall.org/products/snapshot

Overview

Search entry view

• Earliest date of citation across all datasets
• Latest date of citation across all datasets
• A list of the languages for which there are datasets

Search results view

• A link to the generated CSV version of the result set

For each search result, we need:

• The language of the result set where the citation occurs
• Wikipedia article title
• Revision identifier of the Wikipedia page (rev id)
• Wikipedia article time stamp
• Cited article's ID

e.g. https://twitter.com/rmounce/status/994875208019398657

DOIs sometimes have /pdf, /abstract, or other oddities we don't want at the end of the identifier. This should be cleaned in the collation script.