- newsrank.org
- mediarank.*
- inforank.*
- sourcerank.*
- Populate a database with data that measures the ‘information quality’ of a publication
- Rank publications according to the data
- Expose these rankings via a web interface and via a public API
- “I trust the experts” - ie. How many academics contribute to the publication?
- “I value sources” - ie. How many links out of an average article on the publication
- “I don’t want click-funded content” - ie. Do they have a subscription revenue model?
- “I want to avoid dishonest publications” - ie. How many time have their articles been review as “false” on Snopes.com, TrueFact, etc. etc.
- "I want to avoid highly biased sources" - ie. Use https://mediabiasfactcheck.com/
- Scrape articles from a list of publications
- Get the authors from the articles
- Check if authors have:
- Published papers (check Google Scholar)
- Published Books
- (Stretch goal) Check if their published stuff matches the article's topic (use an ML API if one exists)
- Rank publications according to who most often publishes academics
- Manually check the list of publications
- Scrape articles from a list of publications
- Check how many external hyperlinks exist on average
- Note: hyperlinks != sources, but it’ll do for an MVP
- Fact check organisations will more often (I think) fact check people rather than individual articles
- So if you search for people in the
author
table and see whether their name shows up on the fact check websites, we can trace this back to the publications they've published on.
MediaBiasFactCheck.com has a lot of data on this, but they don't seem to provide a public API. Scraping their list-view pages shouldn't be too hard though.