"Most in-demand programming skills" lists are usually surveys of what employers wish they could find, or what course platforms want to sell. This one is different: it ranks skills by how many developers in a 2.5 million profile index actually show them in public work.
Demand in the job-market sense and supply in the public-code sense are not identical — but supply is what your sourcing filters hit first.
Counts are developers with the skill in public GitHub activity, folded through a canonical taxonomy. Data as of August 2026. Browse live numbers on each skill page.
Top skills by indexed developers
| Rank | Skill | Developers |
|---|---|---|
| 1 | JavaScript | 1,336,934 |
| 2 | HTML | 1,202,938 |
| 3 | Python | 1,153,995 |
| 4 | TypeScript | 749,904 |
| 5 | Java | 743,885 |
| 6 | CSS | 684,048 |
| 7 | C++ | 510,023 |
| 8 | Jupyter | 433,818 |
| 9 | C | 387,193 |
| 10 | Bash / Shell | 367,495 |
| 11 | PHP | 360,706 |
| 12 | C# | 298,627 |
| 13 | Go | 255,927 |
| 14 | Ruby | 201,639 |
| 15 | Vue.js | 151,658 |
| 16 | Rust | 127,090 |
| 17 | Kotlin | 123,406 |
| 18 | Docker | 114,850 |
| 19 | Dart | 114,828 |
| 20 | Swift | 105,978 |
How to read this as a hiring manager
Large pool ≠ easy hire. JavaScript and Python searches return more people than you can process. Your problem is ranking and outreach quality.
Smaller pool ≠ "nobody exists." Rust at ~127k is still a real market — but every message has to earn a reply, and geography matters more.
HTML/CSS rank high because they co-occur with almost every web stack. Do not open a "hire HTML developers" req unless you truly mean it; use them as secondary filters.
TypeScript vs JavaScript. TypeScript is smaller but the population is more deliberate (higher median contributions in our data). For serious frontend/Node roles, prefer TypeScript as the primary skill page.
Skills that punch above their survey hype
Go and Rust sit below the Java/Python giants but hire like scarce specialties. Expect longer sourcing cycles and higher comp.
Kotlin and Swift are mobile-gated — pool size is country-sensitive. Check skill × country pages before you assume a global remote search will work.
Docker as a primary skill is a smell. It is usually a secondary filter on top of a language.
Practical sourcing against this list
- Pick one primary skill and at most two secondaries.
- Open the skill hub and note median activity — set filters relative to that median.
- Add country only after the global shortlist is too large or timezone-constrained — where developers live.
- Screen on proof, not on the skill tag alone — GitHub screening.
What surveys get wrong
Surveys overweight what CTOs plan to adopt next year (AI tooling, niche frameworks). Public code overweight what people already ship. For filling a seat this quarter, ship-weighted data is the better sourcing map. For planning a platform migration, surveys and internal architecture matter more than GitHub totals.
For the behavioural side of "who is strong," see what makes a great developer.