Transforming Real Estate Insights: The Role of Data Science in Forecasting Capital Flows
Our latest Active Capital update shares our investment predictions to help you navigate the real estate market, but how do we generate the data? We speak to Dr James Culley, Head of Data Science, to learn about the role of AI and machine learning in formulating our Capital Gravity model forecasts.
1 minute to read
Our Capital Gravity model forms the cornerstone of our Active Capital research, providing cross-border capital flow forecasts across all sector and investor types.
Applying machine learning models built on commercial real estate cross-border investment data*, our forecasts integrate a range of factors, including economic, financial, risk-based, spatial and cultural data. Amongst the economic variables, we have found that long-term government bond yields, central bank interest rates, exports, exchange rate and risks are significant contributors to explanatory power.
The forecasts are then stress-tested and calibrated across our global network of investment experts and regional research departments.
You can hear more details on our approach in the below video.
At Knight Frank, we have been immersed in the realm of data science for over a decade; our team has been dedicated to applying advanced techniques to real-world business challenges, particularly across the dynamic landscape of real estate.
We have used data science methods to automate, support and expand our research intelligence. Whilst we use machine learning and AI in areas such as natural language, processing, content generation, recognition and pattern detection, it is in predictive modelling, including forecasting, that data science really benefit our work on real estate investment.
*Data sources include: RCA, Oxford Economics, CEPII and Damodaran Online.