Financial Inclusion with GenAI

A web-based Generative AI-powered solution for enabling alternative credit scoring and promoting financial literacy among underbanked populations, built for the Salesforce Futureforce AI Challenge 2025.

🏆 Project Overview

This platform leverages Generative AI, explainable ML, and Retrieval-Augmented Generation (RAG) to address financial inclusion. Built as a modern web app with:

• Frontend: React (with Vite and Tailwind CSS)
• Backend: Node.js + Express
• Database: MongoDB

Key Features

• Alternative Credit Scoring: Predicts user credit scores using synthetic data and ML models for users lacking formal credit history.
• Explainable AI: Uses SHAP values to explain scores, with Gemini API for user-friendly explanations.
• Educational Chatbot: Personalized RAG-based chatbot for financial literacy and awareness.
• Scheme Discovery: Workflow for users to discover and apply to eligible NGO or government support schemes.

---

🚀 Demo

Coming soon!

---

🖼️ Architecture

Frontend (React + Vite)  <--->  Backend (Node.js + Express)  <--->  MongoDB
         |                                 |                        |
         |---> RAG chatbot (EduChat)       |---> ML/SHAP API        |
         |---> Credit Score Dashboard      |---> User/Auth API      |
         |---> Scheme Application          |---> Data API           |

---

📦 Installation

1. Clone the repository

git clone https://github.com/Jaisman/Financial_Inclusion.git
cd Financial_Inclusion

2. Setup the backend

cd backend
npm install
# Make sure to update MongoDB connection string in backend/index.js if needed
npm start

3. Setup the flask backend

cd ../flask_backend
pip install -r requirements.txt
python app.py

4. Setup the frontend

cd ../frontend
npm install
npm run dev

4. Access the app

• Frontend runs on http://localhost:5173
• Backend runs on http://localhost:8000
• Flaks Backend runs on http://localhost:5000

---

🛠️ Technologies Used

• Frontend: React, Vite, Tailwind CSS, React Router
• Backend: Node.js, Express, Mongoose, JWT, CORS
• Database: MongoDB Atlas (or local MongoDB)
• AI/ML: SHAP, Gemini API, Gradient Boosting Regressor (ML code in /flask_backend or external service)
• Other: Axios, bcrypt, cookie-parser

---

💡 Solution Highlights

• Synthetic Data Generation: For alternative credit scoring, since real underbanked datasets are hard to find.
• Explainable ML: SHAP values integrated with Gemini API to generate user-friendly explanations.
• RAG Chatbot: EduChat guides users on financial literacy, via a floating button.
• User Workflows: Auth (login/register), profile management, dashboard, credit score insights, scheme application, financial quiz.

---

📁 Repository Structure

.
├── backend/           # Node.js + Express backend (user APIs, MongoDB models, ML integration)
├── frontend/          # React frontend (Vite, Tailwind, EduChat, dashboards, etc.)
├── flask_backend/     # Python backend for ML/SHAP/Gemini logic if needed
├── .vscode/           # Editor/IDE config
├── README.md
├── package.json
└── package-lock.json

---

🔑 Sample Code Highlights

• Frontend Routing: frontend/src/App.jsx
• User Model: backend/models/user.js
• MongoDB Connection: backend/connection.js
• User API Routes: backend/routes/user.js
• Sample Data Insertion: backend/sample.js

---

📝 How to Contribute

1. Fork this repo
2. Create a new branch
3. Commit your changes
4. Open a Pull Request

---

📚 References

• SHAP Documentation
• Gemini API
• Gradient Boosting
• RAG Principles

---