You've seen the GitHub stars. The trending repos. The "10x developer" myths.
We analyzed every metric. Seriously. We counted.
"Do you need a massive army to move fast?"
We calculated the Commits per Contributor ratio for every project.
Commits per active contributor over the last 12 months
The Verdict: Small teams with fewer than 50 contributors can produce up to 1,138 commits per person. Meanwhile, some mega-projects average fewer than 3 commits per contributor. The "Special Forces" of open source are real.
"If a project responds instantly to issues, they probably fix them faster too, right?"
We correlated Response Time (how fast they say "Hello") with Resolution Rate (how often they actually close the issue).
Each dot is a project. Hover to explore.
That's basically zero. Fast bots saying "Thanks for your issue!" doesn't mean the bug gets fixed.
The Verdict: Don't be fooled by fast response times. Many projects use bots that reply instantly but never actually address the underlying issue. Speed ≠ Quality.
"Is the project building a skyscraper or just painting the walls?"
We compare Commit Activity against Codebase Size.
Projects in the upper-left are "maintenance heavy"
| Rank | Project | Commits | Codebase Size | Maintenance Ratio |
|---|
The Verdict: Projects like Model Context Protocol (MCP) appeared with huge activity but small size—classic signs of a new, rapidly iterating standard or heavy refactoring.
"Which projects have huge corporate buy-in but relatively small contributor circles?"
We look for a high Organizational Diversity Ratio (Organizations / Contributors).
Larger dots = higher organizational diversity
The Verdict: These "Hidden Gems" are often the safest bets for enterprise adoption—stable, backed by many, but not chaotic. Think critical infrastructure libraries like Numcodecs or ko.
"Which projects are punching way above their weight (and thus have the highest 'Bus Factor' risk)?"
Projects with ≤50 contributors generating massive commit volumes.
⚠️ Risk: High output from a small group means if one key person leaves, the project could stall.
Total commits from projects with 50 or fewer contributors
| Rank | Project | Total Commits | Linux Foundation |
|---|
The Verdict: These are the "David" projects of the ecosystem. Impressive, but fragile. If you depend on these, make sure there's a succession plan.
"Who is running out of steam?"
We looked at projects with High Productivity Scores that have seen a massive drop in momentum (commits plummeting compared to the previous period).
Red dots = slowing down. Green dots = accelerating.
| Rank | Project | Productivity | Current Commits | Previous Period | Momentum |
|---|
The Verdict: We found alarming drops in projects like Islet and CheriBSD (>97% drop). These teams were running hot but are now stalling. If you depend on these, check on them.
"Are they building new features or just rewriting the same code forever?"
We calculated a Churn Ratio: Commits per Net Line of Code Change.
| Rank | Project | Commits | Code Changes | Churn Ratio |
|---|
The Verdict: Projects like Model Context Protocol (MCP) and EVerest have ratios >2,800. This indicates massive refactoring, stabilization, or non-code work. They are spinning their wheels (or polishing the engine) rather than driving forward.
"Is a low contributor count always bad?"
We segmented our "Hidden Gems" into Libraries and Apps.
Color-coded by type: Libraries vs. Apps vs. Hybrid
The Verdict: This distinction saves us from flagging a perfectly healthy library as "stagnant". Context matters—a low contributor count on a library is often a sign of stability, not neglect.
After analyzing 5,000+ projects, here's what we learned about open source health
Small, focused teams consistently outperform massive communities on a per-contributor basis.
Fast response times don't correlate with actually fixing issues. Correlation: 0.03.
Critical infrastructure often has high corporate backing but few contributors. These are your best bets.
Small teams with massive output are impressive but fragile. Know your dependencies.
Some projects dropped >97% in activity. High productivity teams can burn out fast.
High commit counts with little code growth often means refactoring, not feature development.
LFX Insights API from the Linux Foundation
Last 12 months (rolling window)
Contributors, commits, orgs, codebase size, response times, merge velocity
Python, Pandas, Altair, Marimo notebooks