Reviving CA Intelligence Suite: An AI Accounting Assistant for Smarter Ledger Analysis

This is a submission for the GitHub Finish-Up-A-Thon Challenge

What I Built

I built CA Intelligence Suite, an AI-powered Chartered Accountant assistant that helps analyze company ledger data from CSV or Excel files.

The project combines machine learning, accounting rules, financial dashboards, anomaly detection, compliance checks, and PDF reporting inside a Streamlit web application.

It started as an experimental academic/side project where I wanted to combine AI with real accounting workflows. The original idea was simple: upload ledger data and classify transactions. Over time, I wanted it to become something more useful for CA-style financial review.

Now the application can:

• Upload CSV, XLSX, or XLS ledger files
• Predict transaction categories using a trained neural network
• Classify entries as Expense, Income, Asset, or Liability
• Show confidence scores for AI predictions
• Detect suspicious or unusual transactions
• Calculate income, expenses, profit, GST, assets, liabilities, and financial ratios
• Generate CA insights and compliance warnings
• Display interactive dashboards and visualizations
• Export processed data and generate a professional PDF report

The goal was to turn a partially completed AI accounting experiment into a polished, usable financial analysis tool.

Demo

GitHub Repository: https://github.com/chandrashekharb369/CA
Pdf Report: https://github.com/chandrashekharb369/CA/blob/main/CA_Report_ABC_Private_Limited_2024-25%20(1).pdf

Screenshots:

The main workflow is:

1. Launch the Streamlit app.
2. Upload a company ledger file.
3. Enable AI predictions.
4. Review transaction categories, confidence scores, dashboards, charts, CA insights, and reports.
5. Download the processed CSV or PDF report.

To run locally:

bash
pip install -r requirements.txt
python run_pipeline.py
streamlit run app.py