Page MenuHomePhabricator

Report overall link numbers for Wikipedia Library partners
Closed, DuplicatePublic

Description

The most basic report we need for Wikipedia Library partners is the overall number of links to their sources present on Wikipedia over time. We should be able to present partners with a simple graph showing the change in number over the duration of the partnership, as well as - when possible - before it starts.

We can provide a list of partners and their relevant LinkSearch query(s). We can also give you back-dated link numbers which have been recorded approximately every two weeks.

Things to note here beyond the obvious include that some partners have multiple search queries if they've donated different sets of resources, and we don't have historical data all the way back to the very start for some partners, as we only started collecting link numbers after our partnership started.

T130058 may be a blocking task for some partners; I'm not sure if this is affecting the API or just the on-wiki link search.

Event Timeline

Something to keep in mind is that while link searches work for most of our partners, at least one will require a text search for their publisher name. This is because they provide .pdfs of books and thus no links are added.

Samwalton9-WMF renamed this task from Report overall link number change for Wikipedia Library partners to Report overall link numbers for Wikipedia Library partners.Mar 27 2016, 11:17 PM
Samwalton9-WMF updated the task description. (Show Details)

The main thing to do here is to load up the list of LinkSearch query parameters that would go with each partner into a table on the analytics-store database server, in the staging database. This would be accessible to join with Event Logging data and make the report very easy to write.

After that, it's a simple query, some yaml config, and an easy dashboard config.

Great; I can make the list of queries available when required.