py-air-quality

Measure and analyse air particulate concentration using a Raspberry Pi and a Nova Fitness SDS011 particulate sensor.

Part 1 - Measure & plot air particulate concentration

Follow these instructions to continuously measure and plot the air particulate concentration. The following instructions assumes that you are using a Raspberry Pi running Raspberry Pi OS, and the default user name 'pi'. If your user name is not 'pi', you will have to adjust the file paths in the following code snippets accordingly.

From the bash command line, create a directory, and clone this repository:

mkdir /home/pi/github
cd /home/pi/github
git clone https://github.com/ingo-m/py-air-quality

Create another folder, where measurement data and plots will be saved:

mkdir /home/pi/air_quality

Install dependencies:

sudo apt-get update
sudo apt-get upgrade
sudo apt-get install tmux libatlas-base-dev gfortran cmake python3-tables

Create a virtual python environment, activate the environment, and install the py-air-quality module with all necessary dependencies:

python3 -m venv /home/pi/py_main
source /home/pi/py_main/bin/activate
pip install -r /home/pi/github/py-air-quality/requirements.txt
pip install -e /home/pi/github/py-air-quality

Connect the SDS011 particulate sensor to one of your Raspberry Pi's USB ports.

To continuously measure the concentration of air particulates every 5 minutes, use cron tab. You can read an introduction to cron tab here.

Open crontab from the bash command line:

crontab -e

If this is the first time you opened crontab, you might be asked which editor you would like to use. I would recommend nano.

Scroll to the bottom of the document (by pressing ↓), and paste the following two lines. If you selected the nano text editor in the previous step, you can paste by pressing Ctrl + Shift + v at the same time. The first line will schedule a new measurement every 5 minutes, and the second line will refresh the plots with the results.

*/5 * * * * /home/pi/py_main/bin/python /home/pi/github/py-air-quality/py_air_quality/measurement/measurement.py
*/5 * * * * /home/pi/py_main/bin/python /home/pi/github/py-air-quality/py_air_quality/server/plot_pollution.py

Leave the text editor by pressing Ctrl + x (at the same time), followed by y and Enter.

You can inspect the cron tab (view only) like this:

crontab -l

By default, the measurement data will be written into a text file located at /home/pi/air_quality/baseline_measurement.csv. Every 5 minutes, after each measurement, a new line will be appended. The file contains three columns, separated by ,. The first row is an epoch time stamp (an easy-to use time format, see Wikipedia for more information). The epoch time stamp is often easier to use when analysing data, and can be converted into a human-readable time format when plotting the data.

There are three separate plots showing:

• The air particulate concentration over the last 24 hours (baseline_last_24_h.png)
• The mean air particulate concentration since the measurement started (combined.png)
• The mean air particulate concentration since the measurement started, separately for weekdays (i.e. Monday to Friday, baseline_weekday.png)
• The mean air particulate concentration since the measurement started, separately for weekends (i.e. Saturday and Sunday, baseline_weekend.png)

For more background information, see:

• https://github.com/ikalchev/py-sds011
• https://www.raspberrypi.org/blog/monitor-air-quality-with-a-raspberry-pi/
• https://openschoolsolutions.org/measure-particulate-matter-with-a-raspberry-pi/

Part 2 - Collect data under different conditions

You might want to collect data under different conditions, such as in different locations (e.g. indoors and outdoors). Or you might want to test whether an air filter can reduce the pollutant concentration within a room. The name of the csv file where measurement data is saved can be adjusted in the .env file located at py-air-quality/py_air_quality/internal/.env. For example, you might first collect data inside a room to establish a baseline:

EXPERIMENTAL_CONDITION="baseline"

After a few weeks of measurements, you could place an air filter inside the room, and change the .env file to:

EXPERIMENTAL_CONDITION="with_filter"

After a few more weeks of measurements, you might then compare the results with and without the air filter.

Part 3 - Show plots on other devices (optional)

At this point, you are able to continuously measure the air particulate concentration, and inspect the resulting plots, which are updated automatically. If you would like to be able to easily view the most recent plots on other devices on your local network (another computer, or your phone) in a web browser, follow the next steps.

Please note that the plots will only be accessible to devices logged into the same local network, i.e. via the same router, for security reasons. Making them available over the public internet is possible, but more complicated.

Check your Raspberry Pi's IP address:

hostname -I

This might output a long string of characters; we are only interested in the first block, which looks like this: 123.456.7.890 (your actual number will be different).

Start the server that will make the plots available on your local network like this, after adjusting the IP address with the one from the previous step:

tmux
source /home/pi/py_main/bin/activate
cd /home/pi/github/py-air-quality/py_air_quality/server/
uvicorn server:app --host 123.456.7.890

The first command (tmux) starts a separate terminal session that you can leave without terminating the process (so that the server stays on). The uvicorn starts a server on the Raspberry Pi, so that other devices (on the same local network) can view the air pollution plots.

Exit the tmux session by typing Ctrl + b and d.

You can now view the plots in a normal web browser, from any other device on your local network, at these URLs (adjust the IP to the Raspberry Pi's address from the previous step):

• http://123.456.7.890:8000/last_24_h
• http://123.456.7.890:8000/combined
• http://123.456.7.890:8000/weekday
• http://123.456.7.890:8000/weekend

You can leave the server permanently active, it consumes practically no system resources when running in the background. If you would like to shut down the server at some point, you first have to re-connect to the tmux session:

tmux attach

Now you can terminate the server process by pressing Ctrl + c, and end the tmux session:

exit

If you are not sure whether the server is still running, you can list active tmux session like this:

tmux ls