SEPTEMBER 2023
COELHO Finance
Financial market analytics platform combining Data Science, ML, and AI applied to investment analysis, portfolio tracking, and opportunity identification — running live on Streamlit with free public access.
Executive summary
COELHO Finance is a financial market analytics platform combining Data Science, Machine Learning, and AI to help users make informed investment decisions, track their portfolios, and identify market opportunities. The platform is publicly live at coelhofinance.streamlit.app — free access, no signup, just open and use.
This project was built with deliberate dual purpose: a useful free tool for investors AND my personal sandbox for testing Data Science, ML, and AI frameworks applied to financial market data before bringing them into client production work.
See it deployed
The Streamlit app is the primary live verification — just open it and explore. These 15 slides are the deeper deployment record: the full feature set in operation, the ML pipelines, and the design choices behind the financial data flows.
What the platform actually does
Three pillars sit on the same financial-data substrate:
| Pillar | What it delivers | Why it matters |
|---|---|---|
| Investment decisioning | ML-driven analysis to help users make informed allocation calls | Most retail tools surface stats; this one applies models on top of stats |
| Portfolio tracking | End-to-end ingestion → reconciliation → visualization of user holdings | Closes the loop: from raw data to a dashboard the user can actually read |
| Opportunity identification | Pattern analysis across instruments to surface candidates for further review | Compresses hours of manual scanning into a filtered shortlist |
What this project really was
A sandbox with users. Like Formula 1 Analytics, every time I wanted to evaluate a new framework for time-series, a different plotting backend, or a fresh ML algorithm for financial data, I prototyped it here against real market data before deciding whether it belonged in client work.
The cost was low, the data was clean and public via yfinance, the iteration was fast, and the live Streamlit app meant the testbed was useful — not just throwaway notebooks. Several stack choices for later financial-domain work originated from prototypes that first shipped here.
Stack
- Python — implementation
- Streamlit — public-facing live interface
- pandas + NumPy — data wrangling and time-series ops
- scikit-learn — ML for investment decisioning models
- statsmodels — time-series statistical analysis
- Plotly — interactive financial visualizations
- yfinance — open historical market data source
What this project proves
- Financial markets as a delivered domain — added to Logistics, Real Estate, CV, Cybersecurity, Sports as production domain experience
- End-to-end live product ownership — data ingestion through ML through live UI, all maintained by one engineer
- Sandbox-with-users discipline — the best testbeds for new tooling are the ones that ship to real users; that’s where edge cases surface