DEV Community

Takuto Yuki
Takuto Yuki

Posted on

One Chart That Explains the Power of AI-Driven Development

#ai #productivity #analytics #softwareengineering

What I Did

I wanted to see how much AI coding agents are actually changing the way people work.
So I looked at a group that’s almost certainly using them heavily: Anthropic engineers.

Here’s what I did:

  1. Collected GitHub accounts that mention Anthropic
  2. Pulled their daily contribution counts
  3. Aggregated the data into a weekly chart

The Chart

Commit activity has been rising sharply — especially since spring.
This group is likely among the heaviest users of AI-assisted coding today.

Now imagine this isn’t Anthropic, but your competitors’ developers.
That’s the perspective worth sharing with your team.

How the Chart Was Made

I wrote a small script to generate it:

GitHub logo pppp606 / github-contribution-analyzer

GitHub Users Analysis Tools

This project contains a comprehensive toolset for retrieving GitHub users/organization members and analyzing their contribution data.

📁 File Structure

Main Tools:

  • github_users_enhanced.py - User/Organization search
  • github_batch_analyzer.py - Batch contribution analysis
  • github_group_trends.py - Weekly group trend analysis ⭐Recommended

Support Tools:

  • github_users_gh.py - Basic user search
  • github_visualizer.py - Comprehensive chart generation
  • github_contributions.py - Individual user contribution analysis

Configuration Files:

  • requirements.txt - Python dependencies
  • README.md - Usage instructions (English)
  • README_ja.md - Usage instructions (Japanese)

Scripts

1. User Search (github_users_enhanced.py) - Recommended

Uses GitHub API via gh command with support for multiple retrieval modes.

Features:

  • Supports keyword search, organization all members, and public members retrieval
  • Flexible configuration via command line arguments
  • Reliable and fast using official API
  • Requires gh CLI authentication

Usage:

# Authenticate with GitHub CLI (first time only)
gh auth login
# Keyword search
python3 github_users_enhanced.py --mode search -q "YourCompany"

# Get all
…
Enter fullscreen mode Exit fullscreen mode
View on GitHub

You can point it at any keyword or org, fetch contribution counts, and produce the same style of chart.
It took about 15 minutes to go from idea to visualization.

Notes

  • User list comes from a simple keyword search:

  • Contribution counts don’t directly equal development speed.

    • Still, AI-driven workflows generally push commit activity higher.

  • No ties to Anthropic.

    • I even looked for a Claude Code affiliate program — none exists 😅
    • And yes, sometimes Codex feels smarter.

Takeaway

Put this next to your own team’s contribution chart, and the contrast speaks for itself.

Top comments (0)