A walkthrough by The Data Nerd
The Hidden Dataset That Predicts Fundraises 3 Weeks Early
How GitHub engineering signals give seed and Series A investors an unfair information advantage. And why you've never heard of this approach.
7-minute read. No video, no webinar signup, no fluff.
The core insight
There is one dataset that predicts startup traction before any other signal exists.
It's not AngelList trending. It's not a warm intro. It's not a pitch deck.
It's the commit log.
GitHub is the largest free dataset of real-time engineering activity on the planet. Over 100 million developers. Millions of organizations. Every commit, every pull request, every new contributor, timestamped and public.
When a startup's engineering velocity deviates sharply from its own baseline, something changed inside that company. They hired. They found product-market fit. They're preparing to launch. Or they just raised money and are deploying it.
In our dataset, this pattern preceded 7 out of 10 fundraise announcements by 3 to 6 weeks.
From our live data
Here's what the signal looks like in practice
carlos-emr Healthcare
Accelerating+199%
Commit velocity change
94
Contributors
1,015
Commits (14d)
Open-source EMR system. Classic engineering hiring burst pattern. Something significant is happening inside this organization.
run-llama AI/ML
Accelerating+376%
Commit velocity change
34
Contributors
81
Commits (14d)
LlamaIndex (Series A, backed by a16z). Engineering velocity surged nearly 4x. This is a deploy frequency spike pattern.
These are real numbers from public GitHub data, not hypothetical examples. Browse all 20 sectors →
Why this works
"I already have enough deal flow"
You do. From the same sources as everyone else.
Your network shows you what other investors are already seeing. By the time a warm intro reaches you, the founder has talked to 3 to 5 other investors. The deck is circulating. Terms are forming.
The deals that generate outsized returns are the ones where you arrive before consensus forms. Before the deck exists. Before the company is "hot."
GitHub engineering signals show you what's happening 3 to 6 weeks before the founder decides to fundraise. Your network gets you to the table. This gets you there first.
"GitHub data is too noisy to be useful"
Raw GitHub data is noisy. Commit counts alone tell you nothing. A bot can inflate them.
But we don't look at absolute numbers. We look at acceleration patterns: when a company's engineering velocity deviates sharply from its own baseline.
A startup that goes from 20 commits/week to 60 commits/week while adding 4 new contributors and spinning up infrastructure repos is not having a noisy week. Something changed.
We track acceleration patterns across 100+ startups in 20 sectors. The signal is not the commit. The signal is the change in trajectory.
"Public data can't give me an edge"
Everyone has access to SEC filings. Quant funds still make billions parsing them faster and smarter.
Everyone has access to satellite imagery. Hedge funds use it to count cars in parking lots and predict quarterly earnings.
Right now, zero investor tools package GitHub activity as a deal flow signal. The data is public. The analysis layer didn't exist.
How many investors in your network are monitoring GitHub commit velocity right now? The answer is probably zero. That's the definition of an edge.
The Insider Circle
What you get when you join
Full Dashboard access
100+ startups ranked across 20 sectors. Filter by sector, stage, geography. Updated weekly. (Worth EUR 49/mo alone)
Private investor Telegram group
Direct discussion with other data-driven investors. Signal alerts, deal flow sharing, market reads.
Monthly signal briefing
Written deep-dive on the month's strongest signals: what's accelerating, what it means, and which sectors to watch.
Custom watchlists
Monitor specific companies or sectors. Get notified when something in your portfolio starts accelerating.
API access
Integrate signal data into your own deal flow tools, CRM, or spreadsheets.
Direct access to the founder
Ask questions, request custom analysis, suggest companies to track.
Early access pricing (locks in forever)
€97
€197
/month
Price goes to EUR 197/mo after beta closes
30-day money-back guarantee. Cancel anytime.
Join the Insider CircleOr start with the Dashboard at EUR 9.97/mo
I obsess over signals others ignore. I watched a company's commit graph spike and three weeks later they announced a Series A. The signal was right there the whole time, public, free, updating in real time.
Nobody was reading it. So I built something that would.
- The Data Nerd
Join the Insider Circle →