Click the image below to watch a short (40 second) video introduction to the project
This project is a simple forecasting tool that estimates whether a hypothetical financial institution would meet regulation requirements under various economic scenarios. It focuses on Basel III regulation and uses sample balance sheet inputs, SQL-based tier calculations, and basic scenario forecasting logic.
This project is written in Python using a Jupyter Notebook.
If you have Jupyter Notebook, VS Code, or Anaconda installed:
- Clone the repo:
git clone https://github.com/markizevbigie/capital-forecasting-microproject cd capital_forecasting_microproject - Open the notebook:
- Launch Jupyter Notebook or VS Code
- Open capital_forecasting_microproject.ipynb
- Run all cells.
- All required packages are standard, but see the “Dependencies” section below for exact versions or install the required dependencies with
pip install -r requirements.txt
- All required packages are standard, but see the “Dependencies” section below for exact versions or install the required dependencies with
- SQLite + SQL Magic in Jupyter
- Relational schema with foreign keys and realistic financial modeling
- Basel III ratio checks and tier-specific compliance logic
- Forecast scenario simulation (baseline, mild/severe recession, expansion)
- Use of CTEs, joins, and CASE logic for clear and auditable calculations
business_units: Simulated departments (e.g., Retail Banking, Wealth Management)balance_sheets: CET1, AT1, Tier 2 capital vs. risk-weighted exposureforecast_assumptions: Scenario-based multipliersregulatory_requirements: Basel III minimums per tier
This project demonstrates a working understanding of regulatory capital frameworks and the ability to implement robust SQL logic for real-world financial risk analysis.
- Add real data from 10-K/Pillar 3 filings
- Automate with Python or build out a front-end dashboard
This notebook was developed using the following dependencies:
-
Python 3.11
-
sqlalchemy==1.4.39 ✅ (Note: newer 2.x versions may cause errors)
-
ipython-sql
-
sqlparse