Analysis updated 2026-05-18
Break down why a stock's price moved by computing real returns for each phase instead of guessing.
Generate an interactive chart showing earnings, ratings, and financing events on a price history.
Create a shareable, high resolution price attribution image for social media.
Check whether a claimed stock return story actually matches the computed numbers.
| kkxatm007/price-attribution | 0xallam/my-recipe | 0xhassaan/nn-from-scratch | |
|---|---|---|---|
| Stars | 0 | — | 0 |
| Language | Python | Python | Python |
| Last pushed | — | 2022-11-22 | — |
| Maintenance | — | Dormant | — |
| Setup difficulty | moderate | moderate | moderate |
| Complexity | 3/5 | 2/5 | 4/5 |
| Audience | researcher | general | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires cloning the repo, running an install script, and installing Python dependencies before generating a chart.
Price-attribution is a tool that helps answer the question of why a stock's price moved the way it did, by breaking its price history into phases and tying each phase to real, computed numbers rather than guesses. The core problem it addresses is that people often explain stock moves with hand-typed percentages that turn out to be wrong once actually calculated, or by cherry-picking only the good quarters while ignoring the bad ones that better explain a drop. You give the tool a JSON file describing a stock, a date range, and a set of phases and events you care about, such as earnings reports, analyst rating changes, or financing actions. From that single spec file, the tool produces either an interactive chart you can explore in a web browser, or a high resolution image ready to share on social media. Crucially, the percentage return shown for each phase is calculated directly from real price data, not typed in by hand, so the numbers cannot drift from reality or be adjusted to fit a story. The project is designed to work well with AI coding assistants like Claude Code, but it can also be used as plain command line Python scripts without any AI agent involved. It supports stock markets in China, the United States, Japan, Hong Kong, Taiwan, and South Korea, pulling price data through existing finance data libraries. A notable design choice is what the author calls a playbook: a set of strict rules meant to prevent common mistakes, like crediting the wrong date for a stock move or mixing up a company's fiscal year with the calendar year. These rules are meant to keep an AI assistant honest by forcing it to rely on code that computes numbers rather than letting it guess or make things up. Getting started involves cloning the repository, running an install script, and installing Python dependencies. From there, someone can either ask an AI agent in plain language to build a chart for a specific stock, or write and run a spec file manually using the included Python scripts. The README does not state a license for this project.
A tool that turns a stock's price history into phase-by-phase charts with computed, not guessed, returns tied to real events.
Mainly Python. The stack also includes Python, Plotly, yfinance.
The README shown here does not state a license for this project.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly researcher.
This repo across BitVibe Labs
Verify against the repo before relying on details.