In the previous rounds of testing we focused on two categories of frequently read articles, one set on controversial current events, and the other on long term popular "evergreen" articles.
For future rounds of testing we want to create a small set of articles tagged by topic. We can use the existing topical analysis and/or APIs to define these, but ideally they'd be broad high level topics of common encyclopedic interest: history, culture, sports, nature, etc.
In order to limit the scope and tech risk, we'll need to keep the article set still on the small side (maybe 100 at most?). So I'd suggest taking the top 10-20 most read articles in the last 30 days in 5-10 topics.
- Biological science
- Physical science
- Ancient history
- Modern history
- Visual arts