FAQ

The GitHub Analytics Tool retrieves data from your GitHub repository and uses it to provide you with insights into your software development team's performance using commonly used software development metrics. It assesses your team based on weekly issue completion, PR completion, cycle time, and open vs. closed issues. It compares your team's scores to the top 100 most starred, forked, and active GitHub repositories.

The GitHub Analytics Tool fetches your team’s data from GitHub and compares that data to the top 100 GitHub repositories. For public repositories, the data available is public, so all that is required to input is the repository URL. For private repositories, you will need to sign into GitHub for us to retrieve that information.

Our current top 100 repositories are listed below:

• nestjs/nest
• forem/forem
• dotnet/AspNetCore.Docs
• firezone/firezone
• spring-projects/spring-framework
• square/okhttp
• ionic-team/ionic-framework
• Significant-Gravitas/Auto-GPT
• puppeteer/puppeteer
• koel/koel
• facebook/react-native
• qmk/qmk_firmware
• square/leakcanary
• microsoft/playwright
• jestjs/jest
• rapid7/metasploit-framework
• Homebrew/homebrew-cask
• keras-team/keras
• streamlit/streamlit
• tailwindlabs/tailwindcss
• freeCodeCamp/devdocs
• gogs/gogs
• microsoft/PowerToys
• neovim/neovim
• discourse/discourse
• swc-project/swc
• minio/minio
• animate-css/animate.css
• netdata/netdata
• protocolbuffers/protobuf
• nushell/nushell
• nuxt/nuxt
• tauri-apps/tauri
• plausible/analytics
• MaterialDesignInXAML/MaterialDesignInXamlToolkit
• pallets/flask
• vercel/hyper
• microsoft/terminal
• ohmyzsh/ohmyzsh
• electron/electron
• apache/dubbo
• laravel/framework
• babel/babel
• alacritty/alacritty
• livebook-dev/livebook
• tachiyomiorg/tachiyomi
• denoland/deno
• AvaloniaUI/Avalonia
• ShareX/ShareX
• chatwoot/chatwoot
• etcd-io/etcd
• composer/composer
• rails/rails
• vercel/next.js
• Ryujinx/Ryujinx
• grpc/grpc
• spring-projects/spring-boot
• ant-design/ant-design
• jquery/jquery
• symfony/symfony
• pallets/jinja
• ultralytics/yolov5
• tensorflow/tensorflow
• supabase/realtime
• taosdata/TDengine
• l20n/l20n.js
• microsoft/vscode
• twbs/bootstrap
• elastic/elasticsearch
• apache/kafka
• starship/starship
• guzzle/guzzle
• meilisearch/meilisearch
• django/django
• apple/swift
• Eugeny/tabby
• elixir-ecto/ecto
• vitejs/vite
• google/flexbox-layout
• angular/angular
• Wox-launcher/Wox
• facebook/react
• traefik/traefik
• Homebrew/brew
• supabase/supabase
• ruby/ruby
• elixir-lang/elixir
• D-clock/AndroidSystemUiTraining
• caddyserver/caddy
• bevyengine/bevy
• google/leveldb
• apache/flink
• yewstack/yew
• coder/code-server
• gin-gonic/gin
• nightscout/cgm-remote-monitor
• OpenBB-finance/OpenBBTerminal
• absinthe-graphql/absinthe
• mybatis/mybatis-3
• flutter/flutter

The GitHub Analytics Tool determines the top 100 GitHub repositories by:

1. Gathering the top 100-starred and 100 forked projects.
2. Filtering out any projects that do not use issues or pull requests.
3. Calculating a set of software engineering metrics based on the progress of each project's issues and pull requests.
4. Normalizing each metric into a score between 30 - 100.
5. Calculating an overall grade based on a weighted average of the individual metrics.
6. Selecting the top-ranking projects as the performance baseline.

The GitHub Analytics Tool calculates the following metrics:

• Issue completion (weekly):
  # Issues completed per week over the last 4 weeks.
• PR throughput (weekly):
  A 4 weeks rolling average of the weekly # of PRs merged.
• Issue cycle time:
  # days for an issue to go from created to closed.
• Code review time:
  The # days between when a PR is opened and when it is merged.
• Open/Closed issues:
  The ratio of open to closed issues.
• Open/Closed PRs:
  The ratio of open, to merge, and abandoned pull requests.
• Open issue labels:
  A list of the top 4 labels on open issues (as well as a count of all other ones).

We would like to acknowledge Goodharts Law and conceed that the metrics above can be gamed to improve the metric score without actually making your team more productive.

The GitHub Analytics tool calculates a set of metrics based on recent GitHub issues and pull requests. Those metrics are scored between 30 - 100 and then combined to create an overall grade based on the following metric weights.

Your grade is compared to the top 100 GitHub repositories. Your score is the percentile your team has achieved compared to these repos (e.g., 75th percentile means that your score is better than 75% of the top 100 GitHub repos).

Yes, you can use the GitHub Analytics Tool to assess personal projects, so long as there is a repository, public or private, associated with the project.

Once you have your results, it’s time to do some evaluation. If your results were not what you were hoping for, it might be time to reconsider some of the processes currently blocking your team. Are you spending too much time on the wrong tasks or updating issues? Are too many meetings preventing your team from having time for focused work? We recommend using a tool like Zenhub to continue gaining insights into your team’s most common blockers, how their work flows through pipelines, and what their capacity looks like.

The GitHub analytics tool only uses GitHub metadata on the repository, issues, and pull requests. This includes:

• Repo private flag (true/false)
• State (open, closed, etc.)
• State_reason, created_at, updated_at
• Closed_at, merged_at
• Userid
• Author_association (member/contributor)

No, the GitHub Analytics Tool does not read, store, or persist your source code.

If you are grading a private repository, the GitHub Analytics tool requires read access to the GitHub project data you would like analyzed. It will never read or store your source code. Public repositories do not require any administrative permissions. For more information, check out our privacy policy here.