Fortune Credit: Seasonal Credit Risk Analysis

Project Overview

A data analysis case study exploring repayment patterns in agricultural lending using simulated loan data.

The project examines whether the timing of a farmer's loan relative to the expected harvest is associated with different repayment patterns.

Note: This is an independent case study using a small simulated dataset. It is not based on Fortune Credit's internal data.

Business Question

Could the timing of an agricultural loan relative to the expected harvest be associated with different repayment patterns?

To explore this, I created a Harvest Gap feature representing the number of days between loan disbursement and the expected harvest date.

Key Observations

The dataset contains:

• 10 total loans
• 7 farmer loans
• 2 farmer loans in the 31–90 day group
• 5 farmer loans in the 90+ day group
• 0 farmer loans in the 0–30 day group

In this small dataset:

• 31–90 days: 2/2 loans repaid (100%)
• 90+ days: 2/5 loans repaid (40%)

These are descriptive observations only. The dataset is too small to establish statistical significance or causal relationships.

Analysis

The project uses:

• Python & Pandas for data cleaning and preparation
• SQLite & SQL for analysis and aggregation
• Streamlit for the interactive dashboard
• Matplotlib for visualizations
• OpenPyXL for Excel reporting

The data processing includes validation of:

• Missing and duplicate loan IDs
• Invalid dates
• Negative loan amounts
• Invalid harvest gaps
• Unexpected repayment statuses

Days_to_Harvest is kept as a numeric field, while timing groups such as 31–90 Days are created separately for reporting and visualization.

Repayment outcomes are also separated into Repaid, Late, and Defaulted rather than treating them as the same outcome.

Project Structure

Fortune-Credit-Data-Analysis-/
├── app.py
├── clean_data.py
├── data_processing.py
├── data/
│   └── fortune_harvest_data.csv
├── requirements.txt
└── README.md

Running the Project

1. Clone the repository

git clone https://github.com/toxidity-18/Fortune-Credit-Data-Analysis-.git
cd Fortune-Credit-Data-Analysis-

2. Install dependencies

pip install -r requirements.txt

3. Run the dashboard

streamlit run app.py

4. Generate static reports

python clean_data.py

AI-Assisted Development

AI-assisted development tools were used throughout the project to help with implementation, debugging, understanding unfamiliar concepts, and improving the code.

I reviewed, tested, and modified the generated implementations while working through the analytical and technical issues identified during development.

Limitations

The dataset is intentionally small and simulated. The results should not be used to make real lending decisions.

A larger historical dataset with actual disbursement dates, harvest dates, repayment dates, loan terms, and customer information would be required for more meaningful analysis.

What I Learned

This project helped me strengthen my understanding of:

• Data cleaning and validation
• SQL-based analysis
• Feature engineering
• Financial data analysis
• Data visualization
• Analytical reasoning
• Reproducible project setup
• Communicating findings without overstating the evidence

Links

Live Dashboard: https://fortunecreditriskanalysis.streamlit.app/

GitHub Repository: https://github.com/toxidity-18/Fortune-Credit-Data-Analysis-