🚀 Inspiration
In India, over 300 million individuals operate in the informal economy without access to formal credit systems. While they often manage their finances responsibly, they remain excluded from financial services due to the lack of a CIBIL score or formal credit history.
We were inspired to solve this challenge using technology for inclusion — to bring these “invisible” citizens into the financial mainstream using the power of AI and alternative data.
💡 About the Project
InclusivScore is an AI-powered credit assessment system designed to provide an alternative scoring method for users with no formal credit history. We aimed to empower GroMo partners to onboard such users confidently, increasing financial inclusion and partner earnings simultaneously.
What We Learned
- How to extract insights from non-traditional data like SMS alerts, UPI activity, and mobile usage
- Building trustworthy machine learning models with sparse or noisy data
- Understanding the financial behaviors of underserved users
How We Built It
- Created a pipeline that ingests alternative data (e.g., SMS alerts, UPI patterns)
- Built a machine learning model trained on synthetic behavioral datasets
- Designed a user-friendly dashboard for GroMo partners to access eligibility results and risk scores in real-time
Challenges We Faced
- Handling privacy concerns and ensuring user consent for data access
- Normalizing diverse data formats across devices and users
- Designing a model that balances inclusivity with risk management
🛠️ Built With
- React.js – frontend interface
- Node.js + Express – backend API services
- Python (scikit-learn) – machine learning model
- Firebase – authentication and database
- Twilio API – SMS parsing simulation
- UPI Mock API – simulated transaction data
- Tailwind CSS – modern UI styling
- Vercel – deployment
Built With
- geopandas
- matplotlib
- numpy
- pandas
- re
- scikit-learn
- seaborn
- textblob
- time
- tweepy
- xgboost
Log in or sign up for Devpost to join the conversation.