The buyer wants to buy the property at the best price, the seller wants to maximize returns. So how do you find the right price?
Housing.com, a real estate advertising platform, tries to solve this problem with data and information. Speaking with ETCIO, Sangeet Aggarwal, Product Manager at Housing.com, explained how the company uses data and mathematical models to predict the price of a property for both sale and rental.
“One of the biggest projects we’ve worked on recently involves predicting home values. There’s no proper framework for telling an agent or seller the right price for their property. Since the ticket is high, how do you know what the best price is to make it available in the market? So we started using the data and made it a data science problem,” he said.
Housing.com is working on something that is known worldwide as the Automated Valuation Machine (AVM). The company captures over 50 home parameters such as neighborhood, bedrooms, bathrooms, balcony, size, windows, floor and other bases on which the models predict the best price to list the property .
The company wants to become as accurate as possible for these price predictions. One of the main things that helps models become intelligent is mapping the actual transaction value to the prediction value so the machine can learn.
“Every three months we come up with a new and better version and map out how well we are priced. But the biggest challenge in this journey is the data itself given the complexity of the Indian market, the areas not mapped and the unavailability of real estate records,” he added.
The company has been working on this project for a while now and it still seems to be a work in progress like every AI project. Aggarwal himself mentioned that there will always be a need for iterations and tweaks.
As the company tries to solve this problem, Aggarwal is ready with another problem statement and a possible solution. Following the e-commerce trend, he wanted to transform the real east platform from directed navigation to directed discovery.
“The new trend in online shopping experiences is discovery-driven search and home buying is no different. That’s something we’re exploring. Buying a home is complicated and a lot of emotions come into play. The buyer is always wondering if he bought the best option, so a discovery-driven platform would assure him that he has seen all the options in the market systematically. But right now there’s a question that says, ‘I’m buying this, could there be better options?’ And that’s the problem we’re trying to solve,” Aggarwal concluded.