Using supervised machine learning to create fast, reliable valuations for commercial real estate.

Supervised Machine Learning

Algorithms can outperform human experts in fields where data is rich and deep, allowing for continuous feedback loops and learning models that can adjust and refine over time. Commercial Real Estate used to lack a comprehensive and sufficiently deep database until GeoPhy developed its proprietary data management platform. Using supervised machine learning and neural nets, our models become more accurate every day.

Real time & instant

Getting an appraisal used to be a process that took weeks, from initial assignment to final report. GeoPhy eliminates that lag in information flow with instant results from our AVM (automated valuation model), dynamically update with new market transactions each day.


We measure the accuracy of our models by comparing our valuations to the flow of sales prices in the market as transactions close. Our current model accuracy is a MdAPE (Median Absolute Prediction Error) of 5,85%. This means our median predicted value is just 5,85% from the actual transaction price — twice as accurate as traditional appraisals for commercial real estate.