Hey everyone, built a little NLP real estate data tool. Trying to figure out if some iteration of this thing can provide some actual value, or if it's just a neat little thing and I should kill it and start my next project.
Basically, it takes a real estate question in natural language, translates it to a database query, and returns a result. Here's a demo: https://imgur.com/a/fOzqSRO
The few real estate folks I spoke to astutely pointed out that the same data can be grabbed from places like Redfin's Data Center (and other dashboards) with a few clicks, which makes sense.
Just wondering if anyone here is involved in RE either as a buyer/seller/investor/landlord, etc. and has some feedback on the possibilities of this tool.
Thanks!
Hi Romulus123, This tool looks promising. If you haven't already I would recommend you look into QA models who do exactly what you want to do.
I could help you build this, do you have a twitter handle/e-mail to reach you?
@Enias what are "QA models"?
QA or Question Answering models are AI models that you feed a corpus, in this case RE data, and you can then query the AI for answers.
Gotcha, yea I've looked into those a bit. I know huggingface has some good tools for that.
I'm not entirely sure what sort of value prop that would provide though. Did you have something specific in mind? The tool I built basically just transforms the query into a set of key words (location, dates, etc.) and translates it into a DB query. What you're referring to is obviously a little different. Curious about what you imagine feeding such a model and what sort of value it could provide/use cases.