According to SBA, there are over 30 million small and medium-sized businesses (“SMB”) in the US and these companies drive nearly 50% of total US employment. The fundamental challenge of small business lending is simple: there’s often a lot of work to be done for a relatively small and risky loan. As a result, many banks avoid small business lending and instead focus on consumer products (e.g. personal loans, auto loans, mortgages, etc.) or larger commercial loans (e.g. $5M+ loans to large profitable companies).
The struggles faced by the small business owners in availing a loan inspired us to create an open lending marketplace - 'Communiti'.
What it does
Communiti is an open lending platform where small business owners could easily post their loan needs and obtain bids from various lenders, thus helping them to get the best possible rates. Lenders can sign up and bid on the posted loan requirements. Communiti helps lenders by intelligently predicting the riskiness of their bid proposals and thereby steering them towards profitable investments. Business owners can view the various bids on their loan and accept or decline them. Lenders can modify and rebid to attractive prospects if needed.
How we built it
We made the server side in Express.js and the client application in Ionic. The database used is MongoDB Atlas. We used the SBA Loans dataset for creating our risk predictor. The predictor was built on RandomForest classifier in scikit-learn and the API for the model was built using Flask.
FFDC APIs used - Customer Onboarding(B2B)
Challenges we ran into
Our proxy was replacing the auth token to FFDC. Building the risk predictor API was a bit difficult. (We lost some sleep but managed to do it!) MongoDB not allowing us to call the update methods. :(
Accomplishments that we're proud of
We were doubtful if we would be able to incorporate the Risk Predictor into our application. However we were successfully able to include that as well on time.
We are proud to have built a working app in Ionic when none of us had previous experience with Ionic.
What we learned
We learned about the various authentication methods used in B2B and B2C APIs in FFDC and how to implement a Python model as an API. We also gained knowledge on the Small Business Loans domain in the US.
What's next for Communiti
Deep learning based models could be used in the Risk Predictor, thus improving the accuracy. Loan Forecasting APIs from FFDC could be included to give a forecast of the repayment process to the borrowers.