Build up a set of discriminatory coded language examples
Examine specific realty websites: Chestnut Hill, Maven, etc.
Question: Is there a distinction between realtors and roommates looking to fill a room?
Q: Is the model of discriminatory language going to be based off of user-reported language, or some pre-existing set of language?
Q: Do we have pre-selected sites that we'd like to focus attention on?
Q: How do you account for edge cases such as neighborhoods with a historical ethnic or cultural heritage?
Q: Can we reach out to Craigslist to provide their flagged listings? Examples of abuse/discrimination? (Craigslist has RSS feeds for selected areas. The easiest thing for them is to say "ya, sure, scrape", the easiest for us is an RSS feed.)
Q: Does the Fair Housing Act apply to vacation rentals? We have FlipKey in the area and AirBnB was in the news recently hiring Eric Holder to help which might give us some data sources / ideas
Q: Is it easier to keep track of discrimination by address not owner?