Our big data methodology is based upon 5 steps:
AggregateWe gather housing demographic, economic, social, and other hyperlocal data to build a complete data set accurately reflecting the investment potential of geographic markets.
TransformData is transformed and normalized. We identify any missing values and fill them in.
Identify FeaturesProprietary features reflecting market conditions are created.
Model Selection and Trend AnalysisUsing statistical and machine learning, we’ve created a wide variety of predictive models and algorithms including generalized linear/mix models, neural networks, decision trees, kernel estimation, time series, etc. Those models are assessed against our real-estate expertise.
Testing and Validation
We test the accuracy and validity of our models by applying the algorithms to historical data.