February 19, 2020

635 words 3 mins read

PostHog/posthog

PostHog/posthog

PostHog is developer-friendly, open-source product analytics.

repo name PostHog/posthog
repo link https://github.com/PostHog/posthog
homepage https://posthog.com
language Python
size (curr.) 2146 kB
stars (curr.) 1391
created 2020-01-23
license MIT License

PostHog header

PostHog (Community Edition)

PostHog is open source product analytics, built for developers. Automate the collection of every event on your website or app, with no need to send data to 3rd parties. It’s a 1 click to deploy on your own infrastructure, with full API/SQL access to the underlying data.

Quick start

1-click Heroku deploy:

See PostHog docs for in-depth walk throughs on functionality.

PostHog dashboard screenshot

Join the PostHog Users Slack if you need help, want to chat, or are thinking of a new feature idea.

Features

  • Event-based analytics at a user level - see which users are doing what in your application.
  • Complete control over your data - host it yourself.
  • Automatically capture clicks and page views to do analyze what your users are doing retroactively.
  • Libraries for JS, Python, Ruby, Node, Go + API for anything else.
  • Beautiful graphs, funnels, user cohorts, user paths and dashboards.
  • Super easy deploy using Docker or Heroku.

Event autocapture

Philosophy

Many engineers find it painful to work out how their products are being used. This makes design decisions tough. PostHog solves that.

We also strongly believe 3rd party analytics don’t work anymore in a world of Cookie laws, GDPR, CCPA and lots of other 4 letter acronyms. There should be an alternative to sending all of your users' personal information and usage data to 3rd parties.

PostHog gives you full control over all your users' data, while letting anyone easily perform powerful analytics.

It means you can know who is using your app, how they’re using, and where you lose users in the sign up process.

What’s cool about this?

PostHog is the only product-focused open source analytics library, with an event and user-driven architecture. That means tracking identifiable (where applicable) user behavior, and creating user profiles. We are an open source alternative to Mixpanel, Amplitude or Heap, designed to be more developer friendly.

There are a couple of session-based open source libraries that are nice alternatives to Google Analytics. That’s not what we are focused on.

One-line docker preview

docker run -t -i --rm --publish 8000:8000 -v postgres:/var/lib/postgresql posthog/posthog:preview

This image has everything you need to try out PostHog locally! It will set up a server on http://127.0.0.1:8000.

Deploy to Heroku

Deploy

Production installation

See wiki for production deployment

Development

Running backend (Django)

  1. Make sure you have python 3 installed python3 --version
  2. Make sure you have postgres installed brew install postgres
  3. Start postgres, run brew services start postgresql
  4. Create Database createdb posthog
  5. Navigate into the correct folder cd posthog
  6. Run python3 -m venv env (creates virtual environment in current direction called ‘env’)
  7. Run source env/bin/activate (activates virtual environment)
  8. Run pip install -r requirements.txt. If you have problems with this step (TLS/SSL error), then run ~ brew update && brew upgrade followed by python3 -m pip install --upgrade pip, then retry the requirements.txt install.
  9. Run migrations DEBUG=1 python3 manage.py migrate
  10. Run DEBUG=1 python3 manage.py runserver
  11. Run the tests and frontend

Running backend tests

bin/tests

Running frontend (React)

If at any point, you get “command not found: nvm”, you need to install nvm, then use that to install node.

  1. Make sure you are running Django above in a separate terminal
  2. Now run bin/start-frontend
  3. Optional: If you’re making changes to the editor, you’ll need to do cd frontend && yarn start-editor to watch changes.
  4. To see some data on the frontend, you should go to the http://localhost:8000/demo and play around with it, so you can see some data on dashboard

Open source / Paid

This repo is entirely MIT licensed. We charge for things like teams, permissioning, data lake integrations, and support. Please email hey@posthog.com and we will gladly help with your implementation.

Contributors 🦸

comments powered by Disqus