Loan Review AI Agent

An intelligent AI agent built with OpenAI's GPT-4 API that analyzes loan application data and answers questions about loan applicants using natural language processing.

This project demonstrates how to build an AI agent that can understand and analyze structured loan data. The agent uses OpenAI's GPT-4 model to provide intelligent responses to questions about loan applicants, their demographics, income, and loan approval status.

• Interactive CLI Interface: Ask questions about loan data in natural language
• AI-Powered Analysis: Uses GPT-4 to understand and analyze loan application data
• Data Context Awareness: Provides the AI with dataset structure and metadata
• Real-time Responses: Get instant answers to your queries about loan applicants

Building_an_AI_Agent_using_OpenAI _API/
├── loan_prediction.py      # Main application file with AI agent logic
├── loan_prediction.csv     # Dataset containing loan applicant information
├── requirements.txt        # Python dependencies
└── README.md              # Project documentation

The loan_prediction.csv file contains loan application data with the following columns:

• Loan_ID: Unique identifier for each loan application
• Gender: Applicant's gender (Male/Female)
• Married: Marital status (Yes/No)
• Dependents: Number of dependents (0, 1, 2, 3+)
• Education: Education level (Graduate/Not Graduate)
• Self_Employed: Employment status (Yes/No)
• ApplicantIncome: Primary applicant's income
• CoapplicantIncome: Co-applicant's income
• LoanAmount: Requested loan amount
• Loan_Amount_Term: Loan term in months
• Credit_History: Credit history status (0/1)
• Property_Area: Property location (Urban/Semiurban/Rural)
• Loan_Status: Loan approval status (Y/N)

1. Clone the repository:

   cd "Building_an_AI_Agent_using_OpenAI _API"

2. Install dependencies:

   pip install -r requirements.txt

3. Set up environment variables: Copy the .env.example file to .env and configure your API credentials:

   cp .env.example .env

   Then edit .env with your actual credentials:

   For OpenAI:

   • Set USE_AZURE_OPENAI=false
   • Add your OPENAI_API_KEY

   For Azure OpenAI:

   • Set USE_AZURE_OPENAI=true
   • Add your AZURE_OPENAI_API_KEY, AZURE_OPENAI_ENDPOINT, and AZURE_OPENAI_DEPLOYMENT_NAME

Run the application:

python loan_prediction.py

Once the agent is running, you can ask questions like:

• "How many loan applications are in the dataset?"
• "What is the average income of applicants?"
• "How many loans were approved vs rejected?"
• "What is the typical loan amount requested?"
• "How many applicants have a graduate degree?"
• "What percentage of self-employed applicants got approved?"

Type exit to quit the application.

1. Data Loading: The application reads the loan prediction CSV file using pandas
2. Data Summarization: Creates a summary of the dataset structure (rows, columns, data types)
3. AI Processing: Sends user queries along with dataset context to GPT-4
4. Response Generation: GPT-4 analyzes the query and provides intelligent answers
5. Interactive Loop: Continues to accept and process queries until the user exits

• pandas: Data manipulation and analysis
• openai: OpenAI API client for GPT-4 integration
• python-dotenv: Environment variable management

The agent supports both OpenAI and Azure OpenAI services.

• Model: gpt-4 (configurable via OPENAI_MODEL)
• Authentication: API Key

• Model: Configurable deployment name
• API Version: 2024-02-15-preview (default)
• Authentication: API Key + Endpoint
• Endpoint: Your Azure OpenAI resource endpoint

• Temperature: 0.2 (for more focused, deterministic responses)
• Max Tokens: 500 (controls response length)

• Data Exploration: Quickly understand loan application patterns
• Business Intelligence: Get insights from loan data without writing SQL queries
• Decision Support: Analyze factors affecting loan approvals
• Educational: Learn about AI agents and natural language interfaces

• The AI agent provides insights based on the dataset summary, not direct data analysis
• For production use, consider implementing actual pandas operations for precise calculations
• Ensure your OpenAI API key has sufficient credits before running the application
• The agent's responses are generated by GPT-4 and may occasionally require clarification

• Never commit your .env file or expose your OpenAI API key
• Add .env to your .gitignore file
• Use environment variables for sensitive information

Feel free to enhance this project by:

• Adding actual data analysis capabilities
• Implementing visualization features
• Expanding the dataset
• Improving the AI prompts
• Adding error handling and validation

This project is for educational purposes.

Built with ❤️ using OpenAI GPT-4