Linked Data Publishing On Activitypub
Linked Attention Data Publishing
Proposer: Ronen Tamari
Basic idea
Attention data is an important and interesting signal. Platforms (e.g., Twitter, Google Scholar, Goodreads, Altmetrics) harvest large amounts of attention data, using it to drive content search and recommendation (CSR). However, this attention data remains mostly locked and inaccessible to the public.
We wish to enable users to easily publish linked data related to attention, or their reactions to content (maybe it could be called “linked attention data”? , Discord discussion)
Background
For broad motivation, see paper/video/thread. And more concretely for our project, many researchers already post such “attention information”, e.g., the Tweet below.
This kind of information can be very valuable for driving content search and recommendation but it gets lost since (1) it isn’t easily machine readable (2) it gets posted to walled platforms with data access issues.
Objective
A possible goal for this project could be a helper app that with the click of a button (maybe a browser extension) would (1) produce some kind of machine readable representation appended to the natural language version of the post (2) publish it using ActivityPub (AP) or other related network (3) possibly mirror it to Twitter if so desired.
Another possible client implementation could be a Zotero plug-in. For example, moving an item to a particular folder or adding a particular note/tag to it could trigger the corresponding AP publishing event.
Implementation
Regarding data schema we can possibly use http://Schema.org schemas such as ReactAction, ReviewAction