Unicorn analytics retrieves analytics data from the YouTube Analytics API and Google Analytics API. The unicorn in the story is the unicorn hat HD. Page hits and video views are visualized as a snake crawling over the unicorn display. One snake equals one page load or view.
If you don't have that many views you can set a scaling factor - for example we can retrieve the statistics over the last week; and display them over 5 minutes - then refresh the statistics and play them again for 5 minutes.
Unicorn analytics also supports operating in API mode, with a REST/WS API to control visualizations. This can be paired with the eve.json events shipper to visualize network events.
GIF of me running the unicorn-analytics on a Raspberry PI 3, replaying page loads and youtube views for the last week over 5 minutes. I used the configuration that is checked in here: unicorn-analytics/config
Related project: pi-zero-ethermeter
You need at least one raspberry pi with a unicorn HAT. The raspberry pi needs python 3.5+ to support await/async.
Modules that needs to be installed,
pip install -r requirements.txtFirst run the unicorn in api mode, this starts the ws and http api's.
python -m unicorn.run --server --port 9990
Then optionally run the eve.json shipper for the api mode, supports ws:// and http://.
python -m shipper.eve --file ./eve.json --server ws://localhost:9990 --token <token>
Check the logs for more information and to ensure the application is operational.
For security a token is required to communicate with the API endpoints. The token is
generated on startup in api mode and logged, alternatively configured in config/server.json.
The following modules are used for authentication.
- oauth2client: used to retrieve oauth2 tokens from google.
- httplib2: to update the oauth token when expired.
Note: the oauth token is for google communications. The token configured in config/server.json is just
used for the server api, when running in api mode.
Now is a good time to configure your google APIs, see the section on Configuration.
To start it all up run,
python -m unicorn.run --analytics --youtubeMake sure you have configured everything first 😏
The first time it is launched it will ask you to open your browser and allow third party access. If you don't have a browser available on your PI, please run as follows:
python -m unicorn.run (...) --noauth_local_webserver
And open your browser on another machine.
Perform the following and update config/analytics.json and config/youtube.json
- create a new Google Analytics project, grab the view ID.
- make sure there is a YouTube channel registered on your Google account.
Create a new API project on the google developers console
- developers console
- enable "YouTube Analytics API"
- enable "Analytics API"
Create an OAUTH2 client secret by following these instructions;
Download it as json and place under api_keys/analytics_secret.json and api_keys/youtube_secret.json.
For the youtube analytics API we may only authenticate using a OAUTH2 token. Most of the examples in the docs assume that we are a web application. Some hours later I actually found out that there IS a flow that can be used from the commandline! Scroll to the bottom of the following page:
https://developers.google.com/api-client-library/python/guide/aaa_oauth#the-oauth2client-library
With this we can actually generate oauth2 tokens without requiring a browser to be opened on the PI. If you don't have a browser available in your environment please pass the commandline parameter
--noauth_local_webserverAll contributions welcome. 🐟 💦