Traditionally, investing decisions in real estate have been driven by historical property prices, the area’s quality, and proximity to high-value local features like grocery stores, schools, and parks. However, properties within a region can still vary widely, making accurate predictions challenging.

Data science methods can use large data sets that go beyond the traditional data and bring in each property’s individual characteristics. It can factor in property characteristics and demographics (among other things) to create granular sub-market indices. It can go right down to predicting property returns in specific postal districts.