mjib007

mjib007 / revenue-yoy-backtest

Public

月營收 YoY 回測分析教材

19
30
69% credibility
Found May 19, 2026 at 19 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Jupyter Notebook
AI Summary

This is an educational tool that lets you test trading ideas based on company revenue growth. You pick a Taiwan stock, set a revenue growth threshold (like 20%), and the tool checks what happened historically every time a company's monthly revenue exceeded that level. It simulates buying the stock on the day the revenue was announced and shows you the win rate and average returns for different holding periods. The tool automatically fetches 10 years of revenue and stock price data, then displays colorful charts so you can see the patterns visually. It's designed as a learning resource to help people understand how financial data relates to stock performance.

How It Works

1
💡 You discover a trading experiment

You find a tool that lets you test whether monthly revenue growth predicts stock price movements.

2
📝 You choose a stock to study

You pick a Taiwan stock by its number, like 2330.TW for TSMC or 2317.TW for Foxconn.

3
🎯 You set your revenue growth rule

You decide: 'Show me what happens when monthly revenue grows more than 20% compared to last year.'

4
You pick how long to hold
📅
Single timeframe

Test one holding period like 20 days

📊
Multiple timeframes

Compare 5, 10, 20, and 30 day holding periods together

5
🔍 The tool looks at 10 years of history

It automatically fetches monthly revenue and stock prices from 2015 to 2025 and finds every time your rule was triggered.

6
📈 You see clear results

The tool shows you the win rate and average return for each holding period, plus colorful charts showing when signals fired.

🎉 You understand the pattern

You now know whether strong revenue growth historically led to profitable trades for this stock.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 19 to 19 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is revenue-yoy-backtest?

A Python script that backtests a simple trading strategy: buy a Taiwan stock when its monthly revenue growth (YoY) exceeds a threshold you specify, then hold for a set number of days. It pulls monthly revenue data from the FinMind API, grabs price data via yfinance, and outputs win rates across multiple holding windows (5, 10, 20, 30 days) along with charts showing buy signals and YoY trends. You run it from the command line with natural language like "2330.TW 月營收 20% 持有20天".

Why is it gaining traction?

Taiwan's stock market has a well-known seasonal pattern where revenue beats can trigger short-term momentum. This tool makes it dead simple to test that hypothesis on any TW or TWO ticker without writing a single line of code yourself. The multi-window analysis is the real hook - instead of picking one holding period, you see win rates across 5/10/20/30 days side-by-side, which helps you tune exit timing. Charts are auto-generated, which saves you from cobbling together matplotlib code.

Who should use this?

Retail traders focused on Taiwan stocks who want to validate the "revenue surprise leads to short-term price spike" pattern before risking real capital. Researchers or students building a portfolio of trading ideas. If you trade US stocks, this is not for you - it's built exclusively for Taiwan exchange tickers (.TW and .TWO format).

Verdict

Early-stage learning material, not production-ready tooling. With 19 stars, no test suite, and a README that's just a binary file, the credibility score sits at a concerning 0.7%. The code itself is functional and the logic is sound, but there's no community vetting or documentation beyond the inline comments. Try it as a learning exercise or a starting point you extend yourself, but do not trust it as-is for live trading decisions. If you want something battle-tested, look for alternatives with active maintenance and test coverage.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.