Repo for code and results of models and exploratory data analysis.
For a thorough introduction to the work Kiva does, check out the detailed welcome doc Jonathan Joa, the Data Ambassador for this project, put together.
Kiva has three main aims for this project:
- I want to use loan, lender and borrower data to explore what types of loans and borrowers are prone to partial loan repayments so that kiva can work with borrowers and lenders to restructure repayments when necessary that maximally suit the borrower (their business is successful and repays loan) and lender (they continue to re-lend) so that that microfinance grows and succeeds with new borrowers getting access to finance and grows the lender base by improving repayment rates.
- I want to use loan, lender and borrower data to predict which loans are at risk of low repayment rates so that kiva can improve their loan risk algorithm which assesses loan risk for prospective lenders so that that microfinance grows and sustains with more accurate and symmetric information.
- I want to use loan and borrower data to explore the words, phrases and characteristics most associated with successful borrower profiles/loans so that kiva and its small business partners can coach new borrowers to create the most effective loan profile for getting fully funded so that that microfinance grows, sustains and maximizes borrowers’ access to finance.
Available via Google Drive. For access, please open an issue in this repo requesting access and a Data Ambassador will follow up with a link.