Introduction

A simple set of scripts to discover an Airthings Wave Plus device, fetch its metrics over Bluetooth, and save the results in a Graphite database. From there, you can graph it with something like Grafana.

It can create dashboards like the following:

Dependencies

Install these dependencies on a Raspberry Pi running Ubuntu:

python3 pipenv bluetooth bluez

There could be more dependencies, but I installed them so long ago I forget any additional.

Setup the Python pipenv environment

Create the pipenv environment and install the dependencies. Check out the pipenv documentation for this. This is usually just a pipenv install command.

Discover your Airthings

Next, find your device. There is a simple find script that scans for 5 seconds and prints out any MACs for Airthings devices it finds, filtering out all the Bluetooth junk from the environment (my neighbors have a bunch of Schlage locks and phones broadcasting).

To run the find script, you need root (preserving the environment for Pipenv to function):

sudo -E pipenv run python3 find.py

Expected output (ignore the logger info at the start of the line):

INFO:root:Found a potential Airthings with MAC address: 58:93:d8:ab:cd:ef

Test the collection script

Finally, you can take your MAC and test that everything is working. The help on the collection script is:

pipenv run python3 ./collect.py --help

Usage: collect.py [OPTIONS] MAC GRAPHITE_HOST GRAPHITE_PREFIX

Options:
  --log-level [critical|error|warning|info|debug]
                                  Log level.
  --graphite-port INTEGER         Port for Graphite.
  --graphite-dryrun / --no-graphite-dryrun
                                  Do not actually send to Graphite.
  --help                          Show this message and exit.

To execute a test run without actually sending data to Graphite, you can use the dryrun flag and any Graphite endpoint and prefix. That looks like this:

pipenv run python3 ./collect.py --graphite-dryrun 58:93:d8:ab:cd:ef 127.0.0.1 my.sample.prefix

WARNING:__main__:Dry-run: Not sending humidity=58.5
WARNING:__main__:Dry-run: Not sending light=74
WARNING:__main__:Dry-run: Not sending accel=0
WARNING:__main__:Dry-run: Not sending radon_short_term_average=5
WARNING:__main__:Dry-run: Not sending radon_long_term_average=12
WARNING:__main__:Dry-run: Not sending temperature=19.98
WARNING:__main__:Dry-run: Not sending pressure=1001.44
WARNING:__main__:Dry-run: Not sending carbon_dioxide_level=465.0
WARNING:__main__:Dry-run: Not sending voc_level=67.0

Set up a shell script and cronjob to run the script.

In a real system, you'd create a cronjob to periodically collect your data. I like to use a shell script to make sure I get the pipenv invocation correct:

A sample shell script is included as collect.sample.sh. Edit the variables for MAC, GRAPHITE_HOST, and PREFIX.

Create a cronjob that looks like this with the path to your collect.sh script. Make sure it is executable.

*/5 * * * * /home/ubuntu/src/waveplus/collect.sh