Future Housing Leaders, Leading with Data Science


July 21, 2020
Polygon Research, Inc. (PRI) had the rewarding opportunity to participate in Fannie Mae's Future Housing Leaders (FHL) summer program, which connects students from diverse backgrounds with paid internships in the housing industry. PRI delivered a four-part class titled: Leading with Data Science: How to Stand Out in the Housing Industry.

Our class provided insights into the building blocks and drivers of the housing industry from the perspective of the data it produces. In addition to primers and documentation delivered through our wiki, we built working data science and machine learning models with live data to enable a better understanding of the industry:
Python
We coded a linear regression model in Python to explore the relationship among several variables in loan level secondary market data
1
Preprocessing
We taught preprocessing techniques in both pandas and SQL
2
HMDAVision®
We demonstrated the power of visual data exploration in our own app HMDAVision, built on Qlik Sense, to understand consumer and lender behavior in the primary mortgage market
3
Classficiation
We built a classification engine in Python with scikit-learn to predict voluntary and involuntary prepayments at the individual loan level
4
Predictive
We leveraged the power of instant machine learning with Tangent Works TIM to forecast aggregate loan prepayment rates based on loan level time series data from the last 20 years
5
In summary
We used real data and showed how to solve real industry problems, with tools and skills that will help the 80 class participants add value in their organizations and stand out among their peers.
6
Polygon Research celebrates the current achievements of the Future Housing Leaders class of 2020 and is excited about the leadership they stand to provide to the housing industry!
Made on
Tilda