We are working on a way to figure out how many men vs. how many women are quoted in any given article.
Here are the steps we've brainstormed:
- Get all names from a given article
- Check to see whether these names are in article.
- predict gender related to name w/API (http://namesorts.com/api/)
- Use Freebase to determine gender of people like rappers ( https://www.freebase.com/people/person?schema=)
- Highlight all known names from database in article
- Allow user to un-highlight a name if it is not a person being quoted
- Names that are already in database are marked as male / female if known
- Ask users to highlight additional names that have not been highlighted
- Ask users whether those names are male or female
- Add those names to database
- Have users press button "I'm done."
- Tell % or scores for the article
- Ask is this accurate?
How to Chrome Extension:
load unpacked extension