What do we do?

We take open micro-data (disclosed by lenders or investors, census data, regulatory data, etc.), and we take it through the entire data science pipeline to make sense of it.

We then activate the data through interactive insights delivered via web-based dashboards

Communication
We deliver commercial Apps. We also deliver custom Interactive Dashboards, defined by clients, and we make tools and processes available via APIs and Data Sharing.
Data Mining
We specialize in delivering instant ML through our partnership with TangentWorks. Our Instant ML work is focused on Predictive Classification analysis, Time Series Forecasting, and Anomaly Detection. We build solution for mortgage prepayment speeds, mortgage production forecasts, and fraud detection.
Freddie Mac PMMS
FHFA HPI
Unemployment rate
Permits/Starts (coming)
Data Transformation
Our data scientists spend significant amount of effort understanding the data. They carefully validate and clean data, and invest time in data enrichment/data blending, and parsing data/data munging.
Census geo files for:
Census tract, zip code, neighborhood, county, MSA, state
Domain Knowledge
Our people are industry experts ranging from CIOs to CMBs. We are housing finance experts building cutting edge cloud apps for housing finance practitioners. We know the issues and we know the data.
12+ million loan level disclosure data about the CRA lending activity
5 years of CRA aggregate lending data
FDIC branch location and deposits
326 million statistical records on people (age, ethnicity, sex, race, income, occupation, employment, etc.
130 million statistical records on occupied housing units and households (ownership, relocation, housing structure age and characteristics, employment, etc) from the American Community Survey, Current Population Survey, Annual Social and Economic Supplements.
800+ million Ginnie Mae loan-level data about originations, and loan performance, including delinquency, liquidations, and forbearance.
90+ million loan-level applications, detailing the loan life cycle, applicant demographics, lender information, granular geo information, property characteristics, detailed loan information.

Mortgage data science

We employ cutting edge cloud analytics and instant machine learning to deliver descriptive, inferential, and predictive analysis at scale. Forecasts for the US residential mortgage market traditionally have been made only at the national level. With our technology partners Qlik and Tangent Works, we provide both rich historical analysis as well as time series forecasts to predict what's coming next. With adaptive feature engineering and fully explainable models, we give you a granular view of where the market is going, and why.