Loading...
Anonymous · Synthetic · Cross-channel
GitDealFlow ships behind a synthetic mascot. The Data Nerd is composed of an abstract constellation avatar (seven dots, one for each atomic GitHub signal), a synthetic voice generated by Cartesia Theo, and a deliberately data-first writing style. The founder’s real name, voice, and face never appear — the framework, the data, and the methodology paper do all the talking.
This page is the public character bible. Operators of partner accounts (LPs, syndicate members, affiliate community managers) use it to keep voice consistent when they re-share or repurpose content. The 30-day content batch updates monthly.
@data_nerd
@gitdealflow.datanerd
GitDealFlowDataNerd
tiktok
@data_nerd_signals
GitDealFlow
youtube
@gitdealflow
threads
@gitdealflow.datanerd
bluesky
@datanerd.gitdealflow.com
substack notes
@gitdealflow
Live — actively posting. Reserved — handle secured, account spin-up scheduled. Cross-posting via existing pipeline.
Primary
Avoid
Hard rules
Signal of the Week
30%
One named startup, one signal, one 30-second walkthrough. Always public-data only. Always with the GitHub URL.
“Signal of the week: ShadcnUI's 0.42 commit-velocity ratio. 14d ÷ 90d = 1.62. Above the 1.5 threshold for the third week running. github.com/shadcn-ui/ui — Insights → Pulse to verify.”
Framework Explainer
25%
Single-signal explainer with the 5-min procedure. Designed to be screenshotted/saved.
“How to read a dependents graph in 60 seconds: Insights → Dependency graph → Dependents. Most investors don't know this page exists. It's the cheapest external-adoption proxy.”
Data Point
20%
One number from the panel. Always cited (SSRN id 6606558). Always with implication.
“Startups that closed had a contributor-diversity Gini of 0.34 at month -3. Startups that didn't: 0.61. The shape of the codebase matters more than the size of it.”
Calibration Case
15%
Look back at a recently-funded startup. What did the framework say at month -3?
“Vercel's Series E announced 2026-04. Composite at month -3: 5/6. The only miss: dependents graph (no public OSS). Framework would have flagged this round before the headline.”
Operator Prompt
10%
Direct ask of the audience — 'try this on a startup you know'. Conversion-driving.
“Try the manual procedure on one startup you almost-invested in. Gut prediction first, then the score. Reply with the delta.”
| Channel | Frequency | UTC slots |
|---|---|---|
| Daily — one Signal of the Week + one Data Point + one Operator Prompt across the day | 13:00, 16:30, 19:00 | |
| 3×/week — long-form Signal of the Week + one Calibration Case + one Framework Explainer | 13:00, 17:00 | |
| 3×/week — carousel format. Signal of the Week (Mon), Framework Explainer (Wed), Calibration Case (Fri) | 18:00 | |
| 1×/week — repurpose the Wednesday Instagram carousel as a Facebook post with extended caption | 14:00 | |
| tiktok | 1×/week — 60-second synthetic-voice walkthrough. Reuses the YouTube Short audio. | 19:00 |
| youtube shorts | 1×/week (Wed) — Data Nerd Brief, character-shaped Short via existing tools/video pipeline | — |
| threads bluesky | Mirror of Twitter. Same content, posted via the existing Substack Notes + ATproto pipeline. | 13:00, 16:30 |
primary
vertical
topical
reach
Signal of the week: ShadcnUI's 14d/90d commit-velocity ratio is 1.62 — third consecutive week above 1.5. Public-data only. github.com/shadcn-ui/ui → Insights → Pulse → 1 month, then 3 months. Compute (weekly × 4) ÷ monthly. Composite is 5/6. signals.gitdealflow.com/predicted?utm_source=social&utm_medium=mascot&utm_campaign=batch-2026-05
**Signal of the week:** ShadcnUI's commit-velocity ratio is at 1.62 — the third consecutive week above the 1.5 acceleration threshold. What that means in plain English: the team is shipping ~62% more in the last 14 days than their 90-day baseline. Sustained acceleration over 14+ days is the cleanest leading indicator we track. The procedure to verify yourself (5 minutes): 1. Open github.com/shadcn-ui/ui 2. Click Insights → Pulse → 1 month. Note weekly + monthly commits. 3. Click 3 months. Note quarterly commits. 4. Compute (weekly × 4) ÷ monthly. If > 1.3, accelerating. If > 1.5, sharply accelerating. The full 7-signal composite scores them at 5/6. The only miss: their dependents graph isn't formally exposed because they're a UI library distributed via copy-paste rather than npm. That's a known framework edge case, not a real negative signal. Backtest: composite of 5/6 with sustained 14d acceleration → ~38% chance of a venture round closing within 47 days. Roughly 5× the base rate. Read the live ranking: signals.gitdealflow.com/predicted?utm_source=social&utm_medium=mascot&utm_campaign=batch-2026-05 Methodology paper (CC BY 4.0): ssrn.com/abstract=6606558
Signal of the week 📊 ShadcnUI — commit velocity 1.62 3rd week running above the 1.5 threshold. Composite: 5/6 Read it yourself: insights → pulse → 1 month github.com/shadcn-ui/ui Link in bio for the live ranking. #GitDealFlow #DataNerd #CodeSideSourcing #VentureCapital #AngelInvesting #GitHubSignals #StartupSourcing #AlternativeData
Signal of the week: ShadcnUI is showing a 14d/90d commit-velocity ratio of 1.62 — the third consecutive week above the 1.5 acceleration threshold. The framework backtest: composite of 5/6 with sustained 14-day acceleration → ~38% chance of a venture round closing within 47 days. You can verify the velocity number yourself in 5 minutes using only github.com (no API, no tool). The procedure is in the free 30-day course at signals.gitdealflow.com/challenge?utm_source=social&utm_medium=mascot&utm_campaign=batch-2026-05. Read the live ranking: signals.gitdealflow.com/predicted
threads
Signal of the week: ShadcnUI 14d/90d commit-velocity = 1.62. Third week above 1.5. Public-data only. github.com/shadcn-ui/ui → Insights → Pulse. Composite: 5/6. signals.gitdealflow.com/predicted
bluesky
Signal of the week: ShadcnUI 14d/90d = 1.62. Third week above 1.5. Verify on Insights → Pulse. Composite 5/6. signals.gitdealflow.com/predicted
How to read a dependents graph in 60 seconds 🔍 Most investors don't know GitHub exposes this page. → Open the org's flagship repo → Insights → Dependency graph → Dependents → Count external dependents (not the org's own repos) The cheapest external-adoption proxy that exists. A few hundred external dependents on a dev-tools startup = strong PMF signal regardless of revenue. Edge case: enterprise/private package registries don't show. That's a sign of paid distribution, not weakness. Save this. Run it on the next dev-tools startup that pitches you. #GitDealFlow #DataNerd #CodeSideSourcing #VentureCapital #AngelInvesting #GitHubSignals #StartupSourcing #AlternativeData
How to read a dependents graph in 60 seconds — most investors don't know GitHub exposes this page. → Repo home → Insights → Dependency graph → Dependents → Count external dependents (excl. the org's own repos) 100+ external = strong PMF signal regardless of revenue.
**How to read a dependents graph in 60 seconds.** Most investors I talk to don't know GitHub exposes this page. It's tucked under Insights → Dependency graph → Dependents and it shows you every public repo that depends on this code. It's the cheapest external-adoption proxy that exists. The procedure: 1. Open the org's flagship repo (the most-starred one, or the one in the README). 2. Click Insights → Dependency graph → Dependents. 3. Count the external dependents — repos that aren't part of the same org. 4. For a dev-tools startup, 50+ external dependents is meaningful. 200+ is strong. What the framework filters for: external usage that the founder doesn't have to talk about for the data to exist. Strong PMF signal regardless of revenue. The edge case worth knowing: enterprise teams using private npm or pip registries won't show up. The dependents page being empty isn't a no-signal. It can be a "private distribution" signal — and that's a different kind of moat. The full procedure for all seven signals is in the free 30-day course: signals.gitdealflow.com/challenge?utm_source=social&utm_medium=mascot&utm_campaign=batch-2026-05
Framework explainer: how to read a dependents graph in 60 seconds. Most investors don't know GitHub exposes this page. Insights → Dependency graph → Dependents shows every public repo that depends on this code. It's the cheapest external-adoption proxy that exists. 100+ external dependents on a dev-tools startup = strong product-market-fit signal regardless of revenue. The full 7-signal procedure (free, no card): signals.gitdealflow.com/challenge?utm_source=social&utm_medium=mascot&utm_campaign=batch-2026-05
Data point from the panel: Startups that closed: contributor-diversity Gini = 0.34 at month -3. Startups that did not close: 0.61. The shape of the codebase matters more than the size of it. SSRN id 6606558.
**Data point from the 219-round panel.** Startups that closed a venture round: contributor-diversity Gini coefficient of 0.34 at month -3 before the announce. Startups that did not close: 0.61. Translation: more concentrated codebases (one or two committers doing everything) close fewer rounds than distributed codebases. The mechanism is straightforward. A 4-person codebase with no single committer dominating is funding a real engineering team. A 1-person codebase is funding a salary. The contract value, the dilution math, and the diligence story are all different — even if the headline metrics look similar. This is signal #2 of the seven we publish. The procedure to read it yourself runs in 5 minutes from any GitHub org's Insights → Contributors page. Methodology paper: ssrn.com/abstract=6606558 (CC BY 4.0) Free 30-day course: signals.gitdealflow.com/challenge?utm_source=social&utm_medium=mascot&utm_campaign=batch-2026-05
📈 Data point from the panel: Closed rounds → Gini 0.34 Didn't close → Gini 0.61 The shape of the codebase matters more than the size of it. (SSRN id 6606558) #GitDealFlow #DataNerd #CodeSideSourcing #VentureCapital #AngelInvesting #GitHubSignals #StartupSourcing #AlternativeData
threads
Data point from the panel: Closed: Gini 0.34 Didn't close: Gini 0.61 Shape > size. SSRN 6606558
bluesky
Closed rounds: Gini 0.34 at month -3. Didn't close: 0.61. Shape of the codebase > size of it.
Try this on one startup you almost-invested in: 1. Gut prediction first — score 0-6 from memory. 2. Then run the 5-min composite. 3. Reply with the delta. Most useful artifact you'll produce this week. Procedure: signals.gitdealflow.com/challenge?utm_source=social&utm_medium=mascot&utm_campaign=batch-2026-05
**One operator prompt for your weekend.** Pick a startup you almost-invested in last quarter. Doesn't have to be high-stakes — just one where you remember saying "I'll think about it" and then didn't. Step 1: write down a gut prediction. Score 0-6 from memory, no looking. What's your guess? Step 2: run the 5-minute composite on their GitHub org. Insights → Pulse for commit velocity, Insights → Contributors for diversity, Insights → Dependency graph for external adoption, README history for freshness, Repositories → Newest for platform buildout, Issues + PRs for ratio. Step 3: compute the delta between your gut and the composite. The artifact you produce — a one-line note with org + gut + composite + delta — is the most useful sourcing-process exercise you'll do this week. It calibrates your gut against the framework, in both directions. Sometimes the framework is wrong; sometimes your gut was. Both are useful. If you reply with the delta (DM is fine), we're collecting cases for the next iteration of the curriculum. Procedure: signals.gitdealflow.com/challenge?utm_source=social&utm_medium=mascot&utm_campaign=batch-2026-05
threads
Try this: 1. Pick a startup you almost-invested in. 2. Gut prediction (0-6) first. 3. Run the 5-min composite. 4. Note the delta. Most useful artifact you'll produce this week.
bluesky
One weekend prompt: 1. Pick a startup you almost-invested in. 2. Gut prediction first (0-6). 3. Run 5-min composite. 4. Note the delta.
**Calibration case — a recently-funded round, scored at month -3.** Vercel announced their Series E in 2026-04. We backtested their composite as it would have read 90 days before the announce. Score at month -3: **5/6.** → Commit velocity ratio: 1.41 (above 1.3 threshold) ✅ → Contributor diversity: top contributor 38% of commits, 7+ active contributors ✅ → Dependents graph: 280+ external dependents on Next.js alone ✅ → README freshness: substantive diff at -42 days ✅ → New repo creation: 4 platform repos in last 30 days ✅ → Issue-to-PR ratio: 0.9 — below the 1.5 threshold (the only miss) The miss on issue-to-PR is interesting because it suggests the team was fielding inbound faster than they could ship — the kind of "we need the round to hire" moment that often precedes a growth-stage close. Composite framework would have flagged this round before the headline. ~5× lift over the base rate at month -3. Calibration runs against known rounds are how you build trust in the framework. Run one a week and the score becomes more useful than the headline. Methodology: ssrn.com/abstract=6606558 Free 30-day course: signals.gitdealflow.com/challenge?utm_source=social&utm_medium=mascot&utm_campaign=batch-2026-05
Calibration case: Vercel Series E 2026-04. Composite at month -3: 5/6. Only miss: issue-to-PR (0.9 vs 1.5 threshold). Framework would have flagged this round before the headline. ~5× base rate.
🎯 Calibration case Vercel Series E (April 2026) Backtested at month -3: 5/6 Only miss: issue-to-PR ratio (suggested inbound > shipping speed) Framework would have flagged this round 90 days early. #GitDealFlow #DataNerd #CodeSideSourcing #VentureCapital #AngelInvesting #GitHubSignals #StartupSourcing #AlternativeData
Calibration case — Vercel's Series E announced April 2026, backtested at month -3. The composite read 5/6 ninety days before the announce. The only miss was the issue-to-PR ratio (0.9 vs the 1.5 healthy threshold) which suggested inbound was outpacing shipping — the kind of "we need the round to hire" moment that often precedes a growth-stage close. Calibration runs are how you build trust in the framework. Run one a week. SSRN: ssrn.com/abstract=6606558
Total in current batch: 5 posts. Full 30-day batch generated monthly from /api/v1/signals.json + /api/v1/methodology.json. Programmatic mirror at /api/v1/social-mascot.json.
See also: /data-nerd (mascot hub) · /about/founder (anonymity rule) · /disclosure (founder disclosure)