Related Posts
Do I need to know ratios for FDD case study?
Additional Posts in The Real Estate Bowl
New to Fishbowl?
Download the Fishbowl app to
unlock all discussions on Fishbowl.
unlock all discussions on Fishbowl.
Do I need to know ratios for FDD case study?
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Download the Fishbowl app to unlock all discussions on Fishbowl.
Copy and paste embed code on your site

Scan your QR code to download
Fishbowl app on your mobile

You are describing API’s. You get API keys to query data on Zillow, census data, Red Fin, maybe counties. You search for land over X acres with a school ranking above a 5, below $50,000 an acre, where unemployment is below 3.7%. That’s the power of integration and API’s. AI/ML could take the returned result photos determine if it’s forest, field, desert based on a long training process. Where I would start is a process that works, but is labor intensive. Then try to apply the right tools to automate it. Maybe you only want fields of grass with the query above because you don’t want to clear trees.
I just have you an example of one. Feel free to ask questions and not listen though.
I worked for a company that could manually measure roofs from GIS data they collected. LiDAR data with drones. They paid offshore workers a few bucks an hour to figure out pitch of roof, and sqft of roof. They could provide these reports within an hour. ML/AI could have saved them quite a bit. Combining lidar data with perimeter measurements they could have pattern matched shapes to identify buildings, then used height to determine pitch. That was 4-5 years ago. I can pretty much guarantee they are trying to solve this now, if they haven’t already. Their savings would have been about 10 million a year. You might think no big deal, but it’s valuable data to roof contractors, insurance companies, solar contractors, then you tie it into estimation software.
As for AI, lots of commercial real estate use cases are made around this type of data without requiring it to connect the actual data sources in a data warehouse. Without normalizing data somewhere the results are highly irregular at times.
We have been guiding our clients to build something internally to make certain their IP and data stay internal rather than it being shared within the open data source.