Platform for peer-to-peer lending - the practice of lending money to individuals or businesses through online services, matching lenders with borrowers.
Predict whether lending request will be approved
Small data – 4,681 entries, 260 features.
Data comprised of records and metrics from weblogs, where each entry corresponds to a specific person.
Semantical categories of the features:
- Workplace: experience, address, position demographic: age, home address, family status, education
- Weblogs: time of web request login, logout, and total login duration
- External resources: socio-economic status according to CBS reports (by cross-referencing home addresses)
R programming language
Business Impact and Insights
- Characterizing population and subpopulations of the peer-to-peer lending community
- Business strategy was redirected according to data analysis
- Machine learning model with 97% accuracy to predict if a lending request will be approved