Dataset Viewer
Auto-converted to Parquet Duplicate
country
stringclasses
3 values
metro
stringclasses
94 values
technology
stringclasses
63 values
role
stringclasses
7 values
developers
int64
50
60.3k
public_email_pct
int64
21
63
open_to_work_pct
int64
4
47
France
*
*
*
11,915
31
19
France
*
*
Software engineering
7,631
33
21
France
Paris
*
*
4,423
32
18
France
*
Python
*
2,977
29
17
France
*
*
AI product
2,558
34
22
France
*
TypeScript
*
2,397
30
22
France
*
*
Infrastructure
2,175
33
20
France
*
Node.js
*
2,000
33
19
France
*
*
AI research
1,588
33
20
France
Paris
Python
*
1,275
29
18
France
*
RAG / Retrieval
*
1,155
33
21
France
*
AI Agents
*
1,050
35
22
France
*
LLM Engineering
*
990
35
23
France
*
PHP
*
902
32
21
France
*
C++
*
899
32
19
France
*
Java
*
895
34
17
France
Paris
TypeScript
*
882
30
22
France
*
React
*
843
32
22
France
Lyon
*
*
759
33
16
France
Paris
Node.js
*
751
32
20
France
*
Kubernetes / DevOps
*
719
35
21
France
*
PyTorch / ML
*
621
31
19
France
*
Go / Golang
*
605
34
19
France
*
AWS
*
583
32
22
France
*
Rust
*
546
32
22
France
Toulouse
*
*
524
31
20
France
Paris
RAG / Retrieval
*
484
33
20
France
*
Computer Vision
*
476
31
24
France
Paris
LLM Engineering
*
443
36
23
France
*
Terraform
*
439
31
21
France
Bordeaux
*
*
421
35
15
France
Paris
AI Agents
*
417
36
24
France
Nantes
*
*
378
30
16
France
*
Angular
*
338
30
20
France
Paris
C++
*
320
35
20
France
Lille
*
*
319
33
17
France
Paris
React
*
313
37
23
France
Paris
PyTorch / ML
*
309
33
20
France
*
.NET / C#
*
294
32
16
France
Paris
Java
*
287
38
18
France
Paris
Kubernetes / DevOps
*
285
35
20
France
*
Azure
*
283
31
20
France
*
GenAI Media (Image/Video)
*
266
36
20
France
*
LLM Inference / Serving
*
262
33
25
France
Paris
PHP
*
254
37
19
France
Paris
AWS
*
246
32
24
France
Paris
Computer Vision
*
246
31
24
France
Paris
Go / Golang
*
242
36
13
France
*
Docker
*
237
34
23
France
Montpellier
*
*
235
33
22
France
Grenoble
*
*
232
34
16
France
*
OVH
*
230
34
21
France
Rennes
*
*
226
35
17
France
*
Ruby
*
211
32
18
France
Paris
Rust
*
211
32
19
France
*
Vue.js
*
198
36
23
France
Marseille
*
*
192
33
21
France
Nice
*
*
184
32
20
France
Paris
Terraform
*
184
28
19
France
*
*
Data science
181
40
20
France
*
*
Data engineering
178
37
21
France
Lyon
Python
*
175
31
15
France
*
Hugging Face
*
174
24
18
France
Lyon
TypeScript
*
161
30
14
France
Strasbourg
*
*
153
34
22
France
*
TensorFlow / ML
*
152
41
23
France
*
Kotlin / Android
*
146
46
28
France
Toulouse
Python
*
146
28
18
France
*
GCP
*
144
43
26
France
Lyon
Node.js
*
140
34
16
France
Toulouse
TypeScript
*
120
22
21
France
*
React Native
*
119
27
24
France
*
NLP
*
113
29
21
France
*
Scaleway
*
113
35
18
France
Paris
Azure
*
113
33
12
France
*
JAX
*
112
37
21
France
*
Flutter / Dart
*
111
31
24
France
Paris
LLM Inference / Serving
*
110
35
25
France
*
Swift / iOS
*
103
38
19
France
Bordeaux
Python
*
101
37
22
France
Paris
Hugging Face
*
100
25
19
France
*
Reinforcement Learning
*
98
33
20
France
Paris
GenAI Media (Image/Video)
*
98
37
17
France
*
GraphQL
*
97
43
19
France
Paris
Angular
*
97
36
21
France
*
PostgreSQL
*
92
27
21
France
Nantes
Node.js
*
91
27
18
France
Lyon
React
*
83
31
22
France
*
MongoDB
*
81
26
23
France
Toulouse
Node.js
*
81
36
23
France
Nantes
TypeScript
*
79
30
16
France
Paris
.NET / C#
*
79
30
22
France
Paris
TensorFlow / ML
*
79
44
29
France
Bordeaux
Node.js
*
77
35
18
France
*
Ansible
*
76
33
29
France
Bordeaux
TypeScript
*
76
38
21
France
Paris
Ruby
*
76
39
16
France
Paris
Vue.js
*
76
38
17
France
Grenoble
Python
*
72
33
7
France
Paris
Docker
*
71
41
24
End of preview. Expand in Data Studio

GitHub Developer Index: reachability and availability by city

Aggregated statistics on 85,938 active GitHub developers in the United States, France and the United Kingdom, broken down by metropolitan area, technology and engineering role.

Each row reports how many developers the group contains, what share publish an email address on their GitHub profile, and what share have flagged themselves as available for hire.

Last updated: 2026-10-02. Refreshed monthly.

Files

File Scope
starhunt-developers.csv All three markets
starhunt-developers-us.csv United States
starhunt-developers-fr.csv France
starhunt-developers-uk.csv United Kingdom

Columns

Column Meaning
country United States, France or United Kingdom
metro Metropolitan area, or * for the whole country
technology Technology, or * for all
role Engineering role, or * for all
developers Number of developers in the group
public_email_pct Share with a public email address on their profile
open_to_work_pct Share whose profile is flagged as available for hire

A developer is counted once per group, never once per technology. Rows are a cube: the country total, each city, each technology, each role, and the city-by-technology crossings.

By country

Country Developers Public email Open to work
United States 60,284 39% 22%
United Kingdom 13,739 34% 21%
France 11,915 31% 19%

Largest cities

City Country Developers Public email Open to work
San Francisco United States 9,844 40% 21%
London United Kingdom 7,009 35% 21%
New York United States 5,709 36% 22%
Seattle United States 4,556 39% 18%
Paris France 4,423 32% 18%
Boston United States 2,831 40% 19%
Los Angeles United States 2,257 38% 21%
Chicago United States 1,746 38% 20%
Austin United States 1,711 40% 22%
Denver United States 1,656 39% 21%
Atlanta United States 1,316 37% 22%
Portland United States 1,227 44% 21%
Washington United States 1,024 38% 23%
Dallas United States 885 39% 26%
San Diego United States 858 39% 19%

Method, and what this does not measure

Figures come from public GitHub profiles and repositories, enriched and deduplicated. Locations are free text on GitHub, so they are normalised to a metropolitan area before counting, and suburbs fold into their hub.

Three limits worth stating.

This measures the developers we have indexed, not every developer in a city: a place with fewer rows may simply be less covered.

"Open to work" is GitHub's own self-declared flag, which people set and then forget. Read it as a floor, not a census.

Groups smaller than 50 developers are excluded entirely. Below that, a percentage is rounding noise rather than a measurement.

Privacy

This dataset contains no individual records: no usernames, no email addresses, no profile URLs. Only counts and percentages over groups of at least 50 people.

License and citation

Released under CC BY 4.0. Free to reuse, including commercially, with attribution.

StarHunt, GitHub Developer Index (2026-10-02). https://www.getstarhunt.com/developer-dataset.html

Full method and tables: https://www.getstarhunt.com/developer-dataset.html

Downloads last month
104