Developing tailored geospatial analytics techniques for real estate

This project will investigate the development of new geospatial analytics techniques for several business applications, mostly focused towards real estate. Geospatial data has attracted interest for decades, however, more and more real big data sets are becoming available, partly driven by massive publicly available data. Classical statistical techniques often suffer from computational tractability in these environments. Recent advances in machine learning allow to process huge volumes of data, and ultimately discovering new insights and useful models from it. However, there still exists a very important opportunity to develop tailored techniques that can take the geospatial component of data into account, together with other available information, in an integrated manner, especially for applications in the real estate domain, such as house price prediction, identification of high-growth markets, real estate recommendation, spatial planning, etc. It is the ultimate goal of this project to develop and evaluate such techniques.