Skip to content

Quantitative research analyzing how GitHub activity in public repositories correlates with short-term stock price movements.

Notifications You must be signed in to change notification settings

tomalmog/CommitTrader

Repository files navigation

CommitTrader

CommitTrader is a quantitative research project analyzing whether public GitHub activity from open-source repositories associated with publicly traded companies has any measurable relationship with short-term stock price movements.

Overview

CommitTrader examines how different forms of GitHub activity align with abnormal stock returns around the event date. The project uses event-study methodology to evaluate whether developer-driven signals contain informational value that public markets may respond to.

Research Focus

CommitTrader investigates:

  • Whether open-source releases correspond with observable stock price reactions
  • Whether commit spikes or elevated development activity correlate with abnormal returns
  • Which GitHub event categories show statistically significant effects
  • How effects vary by company size, sector, or repository relevance
  • Which event windows (e.g., −1 to +1, −5 to +5) exhibit the strongest signals

Methodology

CommitTrader uses a structured event-study framework consisting of:

Event Identification

  • Collection of GitHub releases
  • Detection of commit frequency spikes
  • Categorization of repository activity events

Expected Return Models

  • Market Model
  • Mean-Adjusted Model
  • Market-Adjusted Model

Abnormal Return (AR) Calculation

  • AR computed as the difference between observed returns and expected returns

Cumulative Abnormal Returns (CAR)

  • CAR aggregated over configurable windows such as:
    • Short windows: (−1, 0, +1)
    • Medium windows: (−3 to +3)
    • Extended windows: (−5 to +5)

Statistical Testing

  • t-tests
  • Sign tests
  • Wilcoxon signed-rank tests
  • Cross-sectional tests
  • ANOVA for comparing event types

Aggregation and Interpretation

  • Company-level and sector-level comparisons
  • Event-type-specific summaries
  • Identification of consistent patterns or null results

Data Sources

  • GitHub Activity Data
    Public events (releases, commits, spikes) collected from mapped repositories.

  • Market Data
    Daily stock prices and index returns used for modeling expected returns.

  • Company–Repository Mapping
    A maintained CSV linking tickers to relevant open-source repositories.

Outputs

CommitTrader produces:

  • Event-level AR and CAR results
  • Aggregated statistical summaries for each event category
  • Cross-company and cross-sector comparisons
  • Significance test outcomes
  • Visualizations, including:
    • AR and CAR distributions
    • Timeline charts
    • Event-type comparisons

These outputs support evaluating whether open-source development patterns show measurable alignment with stock market behavior.

Scope and Limitations

  • Public GitHub repositories reflect only part of a company's engineering activity.
  • Effects may vary by time period, company, or sector.
  • Results should not be interpreted as trading signals.
  • The project is intended strictly for research and analysis.

Purpose

CommitTrader provides a framework for studying how open-source software ecosystems interact with financial markets. It enables structured, empirical evaluation of whether developer activity produces market-relevant signals and offers a foundation for further academic research in quantitative finance, software engineering analytics, and market microstructure.

About

Quantitative research analyzing how GitHub activity in public repositories correlates with short-term stock price movements.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 2

  •  
  •