Paperbuzz provides a programmatic API showing how often a scholarly article was mentioned around the web.
The primary source of data is the Crossref Event Data API.
Paperbuzz downloads data from the API often, consolidating events into a database.
An example API call used to gather data is:
https://api.eventdata.crossref.org/v1/events?rows=10000&filter=from-collected-date:1990-01-01,until-collected-date:2099-01-01,obj-id:10.1371/journal.pone.0177707 Metadata comes from oaDOI (Unpaywall) and Crossref, with example metadata calls being:
https://api.oadoi.org/v2/10.1371/journal.pone.0177707https://api.crossref.org/works/10.1371/journal.pone.0177707
You may sometimes see a difference in the amount events between Crossref and Paperbuzz, due to the way Paperbuzz handles Wikipedia events. Paperbuzz handles Wikipedia events differently in two ways.
First, if a source is cited multiple times on a single page, then that is recognized as a single event in Paperbuzz. Crossref would count that as multiple independent events.
Second, is that Paperbuzz does not save events that are on a pages marked as old revisions.
The way Crossref handles wikipedia events can be found here: https://www.eventdata.crossref.org/guide/sources/wikipedia/.