CreditEye

Key Features

• Document Upload and Classification: Users can upload required documents such as IDs, proof of income, and bank statements. The system provides options to classify each document type, ensuring accurate processing and analysis.

• Automated Document Analysis: Leveraging computer vision and OCR technology with Tesseract, the system extracts key information from uploaded documents. This includes validating the authenticity of documents and detecting any potential signs of fraud.

• Risk Assessment and Prequalification: The backend, powered by FastAPI and ChatGPT, analyzes the extracted data against predefined rules and criteria. The system evaluates the applicant's risk score and provides an initial prequalification status, helping analysts make informed decisions quickly.

• Dashboard and Search Functionality: A user-friendly dashboard allows analysts to manage and review applications efficiently. The dashboard includes search and filter functionalities, making it easy to find specific applications and focus on those requiring attention.

• Notifications and Account Management: The system includes notification features to keep users informed about the status of their applications. Account management features allow users to update their information and adjust notification preferences.

Technologies Used

• Frontend: Built using modern JavaScript frameworks with a responsive design to ensure accessibility across devices. Components are designed with Tailwind CSS for a clean and consistent UI.

• Backend: Developed with FastAPI, providing a robust and scalable API for handling application logic and data processing. The backend integrates Tesseract for OCR tasks and utilizes ChatGPT for intelligent data evaluation and recommendations.

Purpose

This project is aimed at enhancing the credit approval workflow for banks by reducing the time and resources needed for document verification and applicant assessment. By automating key parts of the process and providing a comprehensive review dashboard, the system helps financial institutions make faster, more accurate credit decisions, ultimately improving customer satisfaction and operational efficiency.

Target Audience

The primary users of this system are banks, financial institutions, and credit unions looking to modernize their credit verification processes. It is tailored for use by credit analysts, loan officers, and other banking professionals involved in the credit approval pipeline.

---

By implementing this system, banks can leverage advanced AI and machine learning technologies to stay ahead of the competition, offering a seamless and efficient credit application experience to their clients.

Installation

Clone the repository

git clone https://github.com/v3gaaa/CreditEye.git

Backend Setup

Step 1: Create a Firebase project:

• In Firebase, under the Build section, enable Firestore and Storage.

Step 2: Obtain Firebase Credentials JSON (Private Key)

• Go to the Firebase Console.
• Navigate to Project Settings.
• Go to Service Accounts.
• Generate a new private key.
• Save the downloaded file in a /config folder inside the /backend directory.

mkdir /config

Step 3: Setup .env file

• Create a .env file in the /backend folder:

touch .env

• Add the following environment variables:

FIREBASE_CREDENTIALS_PATH=/path-to-private-key.json # Inside /config
FIREBASE_STORAGE_BUCKET=your-project-id.appspot.com
DATABASE_URL=https://your-project-id.firebaseio.com/
PORT=8000
OPEN_API_KEY="sk-..."

Step 4: Install Requirements

• Create a virtual environment:

python -m venv venv

• Activate the virtual environment:
  • On Windows:

venv\Scripts\activate

• On Mac/Linux:

source venv/bin/activate

• Install dependencies from requirements.txt

pip install -r requirements.txt

EXTRA: In some cases, you may need to install these two dependencies separately:

• For MacOS:

brew install pdf2image
brew install tesseract

• For Windows/Linux: Install pdf2image and tesseract through your preferred package manager or directly from their websites.

Frontend Setup

Step 1: Navigate to the /frontendrepository:

Go into the frontend folder where the React Vite project is located:

cd /frontend

Step 2: Install dependencies

Install the necessary packages using npm:

npm install

Usage

Starting the FastAPI server

• Navigate to the /backend directory:

cd /backend

• Run the FastAPI server:

uvicorn app.main:app --reload

Starting the Vite App

• Navigate to the /frontend directory:

cd /frontend

• Run the vite app:

npm run dev