README.md

ReguCheck — Regulatory Risk Reporting System

ReguCheck is a compact, end‑to‑end risk data pipeline that mirrors how banks turn raw loan files into a reporting‑ready view: ingest → validate → reconcile → report. It uses public Lending Club data to demonstrate governance checks, control reconciliations, and business‑friendly reporting in one flow.

Live Demo: https://regucheck-risk-engine-v9qextfge3a8chvnywvqtg.streamlit.app/

What this project does

• Automated data ingestion from Kaggle using kagglehub
• Governance & validation checks for completeness, accuracy, and validity
• Reconciliation control against a mock General Ledger (GL)
• Interactive dashboard built with Streamlit and Plotly

Architecture (Pipeline)

1. data_loader.py
   Downloads the dataset, locates the CSV, samples to 10,000 rows for speed, standardizes column names, and outputs staged_loan_data.csv.

2. governance_engine.py
   Applies data quality rules and splits the data into:

   • validated_loans.csv (clean)
   • dq_exceptions.csv (errors with Error_Reason)

3. reconciliation.py
   Compares validated exposure totals by risk rating vs. a mock GL, calculates variance and status, and outputs recon_report.csv.

4. app.py
   Streamlit dashboard with three tabs:

   • Portfolio: KPIs + exposure by risk rating
   • Data Quality: exception records + failure rate
   • Reconciliation: variance report with conditional formatting

Getting started

1) Install dependencies

pip install -r requirements.txt

2) Run the pipeline (in order)

python data_loader.py
python governance_engine.py
python reconciliation.py

3) Launch the dashboard

streamlit run app.py

Dashboard guide (plain English)

• Portfolio tab
  Shows total exposure and weighted average rate. The bar chart displays exposure by Risk Rating (A–G) where A = lowest risk and G = highest risk.

• Data Quality tab
  Shows rows that failed validation and a Data Failure Rate (% of bad rows).

• Reconciliation tab
  Compares data vs. mock GL totals. Variance > 1% is flagged as Investigation Required.

Screenshots

Hosting / Deployment

Can we host on GitHub Pages?

Not directly. GitHub Pages only hosts static sites, while Streamlit apps require a Python server.

Recommended options

• Streamlit Community Cloud (free for public GitHub repos)
• Render, Railway, Fly.io, or Heroku

Streamlit Community Cloud (quickest)

1. Push this repo to GitHub.
2. Go to https://share.streamlit.io
3. Select your repo and set app.py as the entry point.

Note: Kaggle downloads on first run may require you to set Kaggle credentials if the environment doesn’t already have them.

Data notes

Raw CSVs are ignored by .gitignore to keep the repo lightweight. The pipeline will regenerate them locally.