github / innovationgraph

GitHub Innovation Graph

GitHub Innovation Graph

This repo contains structured data files of public activity on GitHub, aggregated by economy on a quarterly basis from 2020 onward.

Through offerings such as the GitHub Innovation Graph, we hope to inform research and public policy that could benefit from data on software development activity globally. We welcome developers, data analysts, researchers, policymakers, and all other interested stakeholders to explore the data, discover insights, and create visualizations, among much more.

The GitHub Innovation Graph provides data on the following areas:

• Git Pushes
• Developers
• Organizations
• Repositories
• Languages
• Licenses
• Topics
• Economy Collaborators

See the datasheet for more information.

Exploring Innovation Graph data

For an overview of the dataset, check out the charts and tables at the GitHub Innovation Graph website.

To dive deeper into the data and run your own analyses, feel free to fork this repo, explore the structured data files using the exploratory data analysis tool of your choice, and share your findings in our Discussions page.

Limitations

The GitHub Innovation Graph dataset contains data on (1) public activity (2) on GitHub (3) aggregated by economy (4) on a quarterly basis. As such, this dataset would not be useful for understanding:

1. private activity;
2. outside of GitHub;
3. at a more granular geographic level than economy; or
4. at a more granular temporal level than quarterly.

Additionally, economies that have fewer developers on GitHub (which generally correlates with the population of an economy) will have less data associated with them in this dataset.

See the datasheet for more information on limitations.

Representativeness of Innovation Graph data

How many economies are included?

We endeavor to publish as much data about public activity on GitHub as possible. However, the number of developers varies considerably by economy, and in some cases we decline to publish specific statistics for economies with fewer than 100 unique developers performing the relevant activity during the specified quarter out of an abundance of caution for developers’ privacy. You can find more information on our methodology in the datasheet.

Below a heatmap shows the count of economies reported for each data file by quarter:

Count of economies by data file by quarter

You can also find the CSV for this heatmap in the data/representativeness_data directory.

Which economies are included?

We aggregate GitHub activity for economies using a definition broader than recognized UN member states. For example, AQ reports activity from developers stationed on Antarctica. Below a heatmap reports the count of data files for each economy by quarter:

You can also find the CSV for this heatmap in the data/representativeness_data directory.

License

This project is released under CC0-1.0.

Maintainers

See CODEOWNERS

Support

See SUPPORT

innovationgraph's Issues

Fixed by @khxu in December https://github.blog/2023-12-07-github-innovation-graph-q2-2023-data-release/

          Fixed by @khxu in December https://github.blog/2023-12-07-github-innovation-graph-q2-2023-data-release/

Originally posted by @mlinksva in #11 (comment)

ASP.NET is not a language

👋

While I am SUPER interested in tracking the popularity of specific framework use, i will say that ASP.NET seems out of place in this graph. For example, we don't list Rails, so I would expect us to also not list ASP.NET.

Thanks!

Consider separating "programming languages" into logical high level groups

Generally I think it's harder for me to wrap my head around the graph we currently have because it's mixing things that are used for different purposes.

For example, Makefile / Shell / Dockerfile / Batchfile / Powershell all seem to be a pretty different logical grouping than Java / C / C# / typescript / ruby etc. It's hard to really make sense of what seeing the plots side by side means . Have you considered showing labels or groups or something that would allow the viewer of the graph to separate on an axis that shows more interesting views? Like "Imperative programming languages"? "shell" languages? Build scripts?

Mostly I'm interested in being able to see things visualized in for categories that are "apples to apples".

Insights into top contributors' employment

Would be great if we could have data for the employers for git pushes.
Assigning the employee to an account may not be trivial, but through the email address or the orgs they belong, it might be feasible.

Would be great to see the top companies contributing to open-source, as well as the top universities (both globally and per region).

Thank you!

P.S.: I'm going to have 250+ students submitting patches to various github projects soon. I would be trilled to see the impact and benchmark it against other universities. They are required to use the university email address in the commits and github account.

How to cite?

This is an awesome new dataset, thanks for releasing it. As it's for research purposes I recommend you explicitly state how you want the dataset to be cited in the README (didn't find it but perhaps I missed it)

Of

Can't export licenses to any compressed format

While trying to export to PNG, JPEG, PDF or SVG in licenses the website loads quite a while then redirects to the following url:

Can someone else reproduce this or is this only on my machine?

Q4 2023 Data Release

Use programming language colors specified in linguist for programming languages chart

Some are very recognizable either due to ubiquity such as Javascript or because the assigned color naturally goes with the language name, like Ruby or Rust https://github.com/github-linguist/linguist/blob/7ca3799b8b5f1acde1dd7a8dfb7ae849d3dfb4cd/lib/linguist/languages.yml#L6139C11-L6139C18

This would make the chart a little more instantly legible for heavy GitHub users (and perhaps users of other tools, I'm mildly curious now how widely those color associations have spread).

[Feature request] "Popular" projects?

This is a really cool project!

I was looking at the India figures and they are astonishing. 14M git pushes. 30M repositories. 12M developers. 450k orgs. In the FOSS communities that we're a part of here in India, a conversation that often comes up is the low number of widely used FOSS projects originating from here. But these numbers are hard to reconcile with that observation (anecdotal, of course) and perhaps indicate a different reality altogether.

It would be great if there was a way to get visibility into "popular" projects from geographies. Stars, forks etc. don't of course directly imply quality, usage, or even popularity, but may serve as an indicator. Perhaps there are other data points that serve as indicators.

Thank you

add F# and Ocaml!

plaase add F# and Ocaml in the diagram/stats

• https://dotnet.microsoft.com/en-us/languages/fsharp
• https://ocaml.org/

Links to the innovation graph in the blog post are broken

What is the working definition of repo in the csv file?

I am looking into the reason that the number of repo declined rapidly in China around Q1 2022. I wonder whether there is a working definition of the repo?

• Would any active repo count as a repo?

• If there is no action on the repo for a given time, would it still count?

• If a repo is archived, would it still count?

Similarly, I wonder whether there is a working definition for organization in the dataset.

Monthly frequency

Thanks a lot for this great project! I was wondering whether there is the possibility to provide this dataset also at monthly frequency? I assume that the underlying data is available. Any feedback is greatly appreciated!

"NOASSERTION" license on License Rankings

https://innovationgraph.github.com/global-metrics/licenses

I think this should be removed or changed to something more obvious like "Unknown". Because it's capitalized I thought "NOASSERTION" might be something about not asserting some copyrights or something and googling "noassertion license" didn't produce immediately obvious results.