Add It Up: Making Sense of the GitHub Octoverse Report

GitHub has just released its annual State of the Octoverse report and it is chock full of data points worthy of analysis. Here are just a few quick takes and a suggestion for future reporting:

1. Non-U.S. Contributors Are Overstated. The report states that nearly 80% of GitHub contributors are not from the United States, but in reality, the data shows that about 20% of open source contributors come from the United States. The distinction is that many GitHub users do not identify their geographic location. For example, earlier this year when The New Stack analyzed the top 7,625 contributors to projects in the LF AI Foundation’s landscape, only 59% had volunteered their location, and that includes the nine people that are located on Planet Earth. When we reviewed the top 100 locations of these contributors, 40% were based in the United States.
2. Huge Dependency Counts Rely on Few Developers. On average, 3.5 million projects are dependent on any of the top 50 npm (Javascript-related) packages. The figure includes dependencies within forks and private repositories, which includes inactive projects and those not primarily written in JavaScript. Only 35 people are direct contributors to the components in each of these packages. Partly due to the unique nature of different programming languages, package managers for Java (Maven), Python (pip), .NET (NuGet) and RubyGems (Ruby) have much fewer dependent projects per direct contributor.
3. Private Repos Outpace Public. Over 44 million repositories were created last year, but the vast majority of those were private. Including previously created projects, 12 million public repositories and forks were pushed to GitHub in the last year. As GitHub, GitLab and Atlassian compete for market share, it is important to realize that much of their revenue is based on managing private repositories.

Source: GitHub’s The State of the Octoverse 2019

We purposefully didn’t include the data about the top projects and languages because that section of the report deserves more detailed comparisons versus previous editions of the State of the Octoverse reports. Before next week’s GitHub Universe conference, we encourage the company to make these older reports easily available so enterprising reporters don’t have to rely on the Internet Archive’s Wayback Machine.

GitLab is a sponsor of The New Stack.

Feature image is from The State of the Octoverse.

Lawrence has generated actionable insights and reports about enterprise IT B2B markets and technology policy issues for almost 25 years. He regularly works with clients to develop and analyze studies about open source ecosystems. In addition to his consulting work,...

Read more from Lawrence E Hecht