4 ms·

Here's the fun regex we use to find some initial candidates in our analysis: val = str.match(/probate|trust.sale|short.sale|auction|cash.offers|all.cash|(s

by rgbrgb 12y ago

joshuaheard 12y ago
Is there a numerical analysis or just looking at the listing agent description? What I look for as a real estate investor is price disparity: a house with a much lower price than others around it. They say you make money when you buy a house, so buying a low-priced house and bringing up to the standards of the neighborhood has been my formula.
- rgbrgb 12y ago
  The regex above is one signal we're using. We also look at relative price per sqft and a number of other factors. We could do a little blogpost that goes in-depth on our algorithm if there's interest. We've also tried some fun stuff like running TFIDF on the descriptions [1]. [1]: http://en.wikipedia.org/wiki/Tf%E2%80%93idf http://en.wikipedia.org/wiki/Tf%E2%80%93idf
  - joshuaheard 12y ago
    This could be a valuable tool. I like the clean website too. I don't see an About page. It would be interesting to see who is doing this and what you plan to do with it.
    - rgbrgb 12y ago
      We're Open Listings (YC W15): https://www.openlistings.co/ https://www.openlistings.co/ We're trying to help people find properties they'd like to buy. We'd be happy to help with the offer and transaction for any home in California, but we're also happy just to provide free tools for home buyers.
  - sandGorgon 12y ago
    could you describe the way you are running TF-IDF on the listings - are you pushing them into Solr/Elasticsearch and then firing a query with the same description to see which scores rank ? Would love to see your query.
rgbrgb 12y ago
I should add that we do something similar for mansions: https://twitter.com/openmansions https://twitter.com/openmansions And here's the full firehose of shitty listings: https://twitter.com/shittylistings https://twitter.com/shittylistings
gcb0 12y ago
i tough you were more scientific. e.g. looking for lower $/sqf in the same area.
- Kiro 12y ago
  Me too but to be honest I think this is way better. It actually got me much more interested in the service than if it was just about number fiddling.

by rgbrgb 12y ago

joshuaheard 12y ago

Is there a numerical analysis or just looking at the listing agent description? What I look for as a real estate investor is price disparity: a house with a much lower price than others around it. They say you make money when you buy a house, so buying a low-priced house and bringing up to the standards of the neighborhood has been my formula.

rgbrgb 12y ago
The regex above is one signal we're using. We also look at relative price per sqft and a number of other factors. We could do a little blogpost that goes in-depth on our algorithm if there's interest. We've also tried some fun stuff like running TFIDF on the descriptions [1]. [1]: http://en.wikipedia.org/wiki/Tf%E2%80%93idf http://en.wikipedia.org/wiki/Tf%E2%80%93idf
- joshuaheard 12y ago
  This could be a valuable tool. I like the clean website too. I don't see an About page. It would be interesting to see who is doing this and what you plan to do with it.
  - rgbrgb 12y ago
    We're Open Listings (YC W15): https://www.openlistings.co/ https://www.openlistings.co/ We're trying to help people find properties they'd like to buy. We'd be happy to help with the offer and transaction for any home in California, but we're also happy just to provide free tools for home buyers.
- sandGorgon 12y ago
  could you describe the way you are running TF-IDF on the listings - are you pushing them into Solr/Elasticsearch and then firing a query with the same description to see which scores rank ? Would love to see your query.

rgbrgb 12y ago

I should add that we do something similar for mansions: https://twitter.com/openmansions https://twitter.com/openmansions And here's the full firehose of shitty listings: https://twitter.com/shittylistings https://twitter.com/shittylistings

gcb0 12y ago

i tough you were more scientific. e.g. looking for lower $/sqf in the same area.

Kiro 12y ago
Me too but to be honest I think this is way better. It actually got me much more interested in the service than if it was just about number fiddling.