About stock_screener

Picking stocks through various screening methods. Focus on Northern Europe.

l

Published by

README.md

Nordic Stock Screener

A stock screener for Nordic exchanges (Stockholm, Copenhagen, Helsinki, Oslo), running as a weekly GitHub Actions job that produces a static website with an accompanying SQLite database.

Live site: lseffer.github.io/stock_screener

Tech stack: Python 3.12 | SQLite | Polars | React + TypeScript + TanStack Table | Vite

Screening methods:

Piotroski F-Score - Financial health evaluation (9 metrics, score 0-9)
Magic Formula - Quality + Value investing (ROIC x EV/EBITDA)
NCAV - Net Current Asset Value (Graham deep value)

Architecture

The screener runs as a scheduled GitHub Actions workflow that:

Discovers stock listings from Nordic exchanges via the yfinance screener API
Fetches financial data (prices, income statements, balance sheets, cash flows) via yfinance
Computes Piotroski F-Score and Magic Formula screening scores using Polars
Stores all data in a SQLite database (incremental — only updates stale records)
Generates a static website with the screening results
Deploys to GitHub Pages

The workflow runs weekly (Saturday 06:00 UTC) and can be triggered manually.

Local development

pip install -r requirements.txt

# Run full pipeline (ETL + site generation)
python generate_site.py

# Skip ETL, just regenerate site from existing database
python generate_site.py --skip-etl

# Run ETL only (populate database, no site)
python generate_site.py --etl-only

Output:

_site/ - Static website (HTML, CSS, JS, data.json)
_site/stocks.db - SQLite database with all raw + computed data
stocks.db - Working database

Frontend development

# Build the frontend (one-time or when frontend changes)
(cd web/app && npm ci && npm run build)

# For live reload during frontend development:
# 1. Generate data first
python generate_site.py --skip-etl

# 2. Copy data into the dev server's public dir and start Vite
cp _site/data.json web/app/public/data.json
cd web/app && npm run dev

Environment variables

All optional:

STOCK_SCREENER_DB — SQLite path (default: stocks.db)
STOCK_SCREENER_OUTPUT — site output dir (default: _site)

Data sources

All data is fetched via yfinance:

Stock listings: Discovered using the yfinance screener API (EquityQuery by exchange) for Stockholm, Copenhagen, Helsinki, and Oslo
Financial data: Prices, income statements, balance sheets, and cash flows via yfinance.Ticker

Running tests

python -m unittest discover tests/ -v

stock_screener