Hacker News Reader

Comment by alyx | Hacker News Reader

Anybody know of any good sources of MLS data for programmatic consumption (even if not free)?

There aren't any. All the regional MLS' are separate entities and make it as annoying/difficult/expensive as possible to access their data programatically. Think having to be a licensed broker, then paying for access to download a daily CSV with all active listings.

Who'd have thought that an industry of middlemen would make it hard to reduce friction in their marketplace...

(Source: building websites with MLS integration for real estate offices)

There are, ListHub, unless you don't think it qualifies as a "good" source? We've been happy enough with it, but we're only on the rentals side of things.

Without having direct access to an MLS (ie, being a brokerage), you're pretty much SOL for getting access to decent listing data.

If you DO have access, and you're looking for a solution to use the MLS data via a more simple API - SimplyRETS (https://simplyrets.com) can help with that :) (Disclaimer: I'm a cofounder).

Past that, ListHub, as another user mentioned, is probably the best bet if you're not a member of an MLS - but it's not free.

I bet Zillow would love to know, too!

Yep. If it were easy to have the MLS data, they would have it. (Disclaimer: work for Trulia/Zillow on exactly this problem.)

See this post I wrote a few years ago: http://geekestateblog.com/real-estate-tech-question-getting-...

Still accurate today.

You could try Listhub, but they charge. Any other source would most likely not be legally obtained.

Listhub has its own problems -- specifically when MLS'es update data, Listhub has to ensure that all of the applications that use its data are updating their information as well. Listhub is also one HUGE xml file that is a pain in the ass to process. (We explored with a python based parser that chunked the file up into smaller listings and then imported that into Elasticsearch; but ultimately gave up and resorted into a C based parser that threw everything into MongoDB.

I'm curious, did you look into streaming XML parsing, StAX-style? It's not hard to process infinitely-long files with constant memory: http://stackoverflow.com/questions/9809469/python-sax-to-lxm...

Reply on news.ycombinator.com