Which programming languages are people searching for?
A live ranking of programming languages by real Google search demand, worldwide and by country, updated every month. As of May 2026, Python leads with 38.2% of tracked demand, ahead of Java. As a category it is concentrated (HHI 0.219). Search interest is the closest public proxy for what is actually being used, so the index shows the pecking order and, more usefully, what is rising and what is fading.
| # | language | ↓ Share | Relative | 12-mo trend | MoM | YoY |
|---|---|---|---|---|---|---|
| 1 | PythonGeneral purpose | 38.2% | ▼ -4.4% | ▼ -7.7% | ||
| 2 | JavaGeneral purpose | 20.6% | ▼ -8.4% | ▼ -21.6% | ||
| 3 | JavaScriptWeb | 12.1% | · -0.6% | ▼ -20.6% | ||
| 4 | SQLData | 9.2% | ▼ -5.8% | ▼ -13.2% | ||
| 5 | PHPWeb | 4.9% | ▼ -5.4% | ▼ -17.2% | ||
| 6 | C++Systems | 4.8% | ▼ -8.3% | · +0.4% | ||
| 7 | TypeScriptWeb | 2.6% | ▼ -2.2% | ▼ -22.2% | ||
| 8 | C#General purpose | 2.2% | ▼ -6.5% | ▼ -24.4% | ||
| 9 | KotlinJVM / mobile | 2% | ▲ +9.6% | ▼ -12.3% | ||
| 10 | GoSystems | 1.7% | ▼ -8.4% | ▼ -23.4% | ||
| 11 | RustSystems | 1% | ▲ +2.1% | ▼ -14.2% | ||
| 12 | RData / stats | 0.6% | ▼ -22.1% | ▼ -30.6% | ||
| 13 | RubyWeb | 0.1% | ▼ -27.7% | ▼ -28.2% | ||
| 14 | SwiftApple | 0% | ▼ -5% | ▼ -29.3% |
Open dataset
Every monthly snapshot is published as open CSV and JSON, free to reuse with attribution. The repository is a versioned, citable record of how demand shifts over time.
Method
Figures are average monthly Google searches for each language, from Google Keyword Planner. Volumes are reported in bands (Google rounds them), so treat them as directional, not exact. Within a country the numbers are comparable; the Share view normalizes each item against the whole category so you can compare countries of very different sizes.
One head term per item. Search demand is nested: “claude ai”, “claude.ai” and “claude code” all sit inside “claude”, so the head term already captures them and adding the variants would double-count. We therefore track exactly one keyword per item. It is a deliberate choice: it can undercount longer-tail searches, but keeps everything on the same footing and avoids double counting. For names that are also a common word (Gemini the zodiac sign, Swift the singer, Go the verb) we use a disambiguated term; unambiguous names use the bare word. The single term for each item is listed below, so the method is fully auditable. Python currently leads worldwide, ahead of Java. Data through May 2026.
| language | Kind | Head search term |
|---|---|---|
| Python | General purpose | python |
| JavaScript | Web | javascript |
| Java | General purpose | java |
| TypeScript | Web | typescript |
| C++ | Systems | c++ |
| C# | General purpose | c# |
| PHP | Web | php |
| Go | Systems | golang |
| Rust | Systems | rust programming |
| Kotlin | JVM / mobile | kotlin |
| Swift | Apple | swift programming |
| Ruby | Web | ruby programming |
| SQL | Data | sql |
| R | Data / stats | r programming |
FAQ
Common questions about how the index is built.
How is programming languages popularity measured here?
We use average monthly Google search demand for each language, pulled from the Google Keyword Planner historical-metrics API. Search demand is nested, so we count exactly one head term per item: "claude" already contains "claude ai", "claude.ai" and "claude code", so counting only the head term avoids double-counting. This can undercount an item's longer-tail searches, but keeps everything comparable. Names that are also a common word use a disambiguated term. Within a country, volumes are directly comparable; the "Share" view normalizes each item against the whole category so countries of different sizes can be compared fairly.
How often is the index updated?
Once a month. Keyword Planner only exposes a rolling 12-month window, so we snapshot every month and keep the history, which is how the index can show longer-term rises and declines that Google itself does not expose.
Which programming languages are tracked?
The current index covers 14: Python, JavaScript, Java, TypeScript, C++, C#, PHP, Go, Rust, Kotlin, Swift, Ruby, SQL, R.
Can I download or cite the data?
Yes. The full dataset (monthly time series per country plus a ranked summary) is published as open CSV and JSON on GitHub, free to reuse with attribution. Each monthly snapshot is committed, so the repository doubles as a citable, versioned changelog.