Ep. 54: Zillow analytics boss on making better predictions thanks to data and deep learning

In this episode of the ARCHITECHT Show, Zillow Chief Analytics Officer Stan Humphries discusses about a wide range of topics, including his unique path from policy scholar to web analytics exec, and how an explosion data helped Zillow grow into the business it is today. Humphries, who helped create the company's Zestimate tool for predicting home values, also shares some insights into how deep learning -- and computer vision, in particular -- helps Zillow further improve its accuracy. It does this in part by analyzing images to extract richer information (e.g., how well a home is maintained, price of appliances, etc.) than simply relying on square footage, neighborhood and standard home-buying data. This episode is sponsored by Datadog and MongoDB.