Predicting the Best Use of Land with Machine Learning & AI
This article, an interview with Bryan Copley, co-founder and CEO at CityBldr, originally appeared in the October 2016 Housing News Report newsletter published by ATTOM Data Solutions. For a free subscription to the award-winning Housing News Report, contact email@example.com.
What is your elevator pitch for CityBldr?
We’re a marketplace for development sites. The backbone of our marketplace is artificial intelligence which determines the best use of every parcel of land.
How do you and your customers utilize ATTOM Data?
If you are a builder or developer we show you the best development sites in any search area, whether or not they’re listed for sale, and give you access to thousands of exclusive owner-registered properties. If you’re a property owner, you type in your address and we tell you if your property is more valuable in its current use or marketed on our platform as a development site to hundreds of developers.
We use machine learning to analyze big property data (ATTOM Data is our largest data set, along with 16 other data sources) that funnel into an algorithm to determine the best use.
The problem with existing property valuation methodologies is that before you understand what a property is worth you need to evaluate not only what it is today but what it could be. This is how developers look at land, and how our algorithm does as well. That’s exactly the problem we’re solving with CityBldr. Is a property more valuable as its current use, or as another permitted use? Then we create a marketplace for buyers and sellers who want info on development sites.
Good use case is a valuable redevelopment property in Capitol Hill, a hot Seattle neighborhood. The owner would realize an additional $1 million in value that they would never know about by going to Zillow. CityBldr tells the owner their property is most valuable to a developer – much more valuable. That’s life changing information that property owners should have before making the decision to sell.
What’s really exciting is how much the algorithm has improved since we released the beta early this year. Due to the nature of machine learning, each time ATTOM feed us a new batch of data at midnight, that course-corrects our algorithm and changes the weighting of our arbitrage index.
How is the marketplace responding to your products/services?
We launched a premium product in May for developers. We now have 300 signed up. In June we launched the property online site and 3,500 have signed up. This is all just in King County, Washington. Last year we launched a consumer-oriented product called Everyhome. We thought, “we’ll just put every property in the country up for sale.”
What we discovered is that it wasn’t consumer buyers that were nuts about the capability to bid on every property – it was developers, builders, investors, etc.
Smart developers are using our data to become super developers, anticipating market changes to get ahead of changes in the market and spot arbitrage.
The specific sweet spot for CityBldr has been multifamily developers and builders. These developers have gone through and vetted our product, and what we’ve found we’re particularly good at is identifying sites that are well-suited for multi-family development. That’s important because there’s a huge need for more housing in the markets we’ve moved into – Seattle, SF, LA, San Diego, and Portland. Our data is also being used by several investors, property flippers, to identify good opportunities that might not be listed on the market.
We’ve had almost 15,000 pro-active offers on off-market properties go out through our site.
There are other cool use cases though: we’re working with cities to identify arbitrage and underutilized sites. We are partnering with non-profit affordable housing developers to find affordable housing development sites. We are working with a large trust to identify keystone sites to promote social good. It really helps to have data to make that decision. An additional upside to all of this is that by improving properties to their best use the byproduct is more functional communities, happier residents, better businesses and ultimately better cities.
Why did you decide to use ATTOM Data as your real estate data partner?
A lot of it had to do with the relationship we had with Matt (our sales representative). Matt gave us a lot of help in determining who would be the right fit for us. He was very forthright. The price seemed good. It was better or comparable with the rest of the market. And the fact that ATTOM is pro-entrepreneur was important to me as well.
What has been your experience with the real estate data delivery?
Been perfect. The data is updated every day, and always arrives on time. Predictably good. Every once in a while we have to tap you guys for something we’re missing, but when we do you are super-responsive in getting it to us.
You get back to us really fast, sometimes outside of business hours.
What has been your experience with the data quality?
It’s been predictable, it’s been streamlined. ATTOM has done a good job of standardizing the data.
What has been your experience with customer service?
It’s been great. I couldn’t ask for anything more on responsiveness and willingness to work with us as a startup.
Please contact us if you have questions about the underlying data referenced in this article, or would like to have access to that data in the form of custom reports, API or bulk files.