I have never given much tought to the quality of indoor air. My first introduction to the monitoring of indoor air was with Amazon Smart Air Quality Monitor, which absolutely sucked. For what it was worth, it was useful and inticed me to explore more. The information provided by the air quality monitors is useful because poor air quality can lead to health issues and lower productivity. Personally for me, I found the Carbon Dioxide measure to be the most important (which, unfortunately, the monitor from Amazon did not provide). It increases noticebly when I feel anxious and do exercise which gives a hint to increase ventiliation.

The Problem

Firstly, why did I even decide to make an indoor air quality monitor? Are there no solutions on the market? The answer is: for the fun of it. More generally, the solutions on the market are not as convinient and customizable. Also, I hate all the awful companion apps that modern products use. No, I do not want a shitty app (that is also somehow a blackhole of everything, social media, store, etc.) that requires me to sell my soul in order to use it. Finally, I always had a fascination with how things work (the amount of things I broke in my childhood to find out . . .). With this project, I have learned interesting things. It reminds me of the quote from the Prelude to Foundation by Isaac Asimov,

The sky was almost cloudless, but it was a pale blue, as though wrapped in a high thin mist. That, thought Seldon, was a good touch, but suddenly he missed the sun itself. No one on Trantor saw the planet’s sun unless he or she went Upperside and even then only when the natural cloud layer broke.

Did native Trantorians miss the sun? Did they give it any thought? When one of them visited another world where a natural sun was in view, did he or she stare, half-blinded, at it with awe?

Why, he wondered, did so many people spend their lives not trying to find answers to questions–not even thinking of questions to begin with? Was there anything more exciting in life than seeking asnwers?

The Solution

Implementation

To begin with, I used Raspberry Pi Zero 2 W for this project which is capable of running the Linux operating system and has onboard Wi-Fi. This is great because running an entire OS simplifies many things and saves a lot of time on implementation. This way, I do not have to write my own server, instead using nginx to provide static pages and API. Though, I most likely will write my own server in C using UNIX sockets to make it more portable (nginx is a pain to setup; it would be great to just compile and run). The very earliest versions of this project used an Arduino Nano, display, and buttons to cycle parameters. However, as you can see from the pictures, it is a lot more convenient to funnel that information over to devices for greater convenience.

For the sensor, I have used an old SEN55 sensor that provides with termperature, humidity, particulate matter, VOC and NOx. On top of that, I really wanted to track CO₂ as it is an extremely important measure for the health. The newer versions of SEN6X sensors, support Carbon Dioxide in one package. But, at the time of making the project, unfortunately, I had to buy a seperate sensor SCD40 to track CO₂. Conveniently, the manufacturer provides Arduino and Raspberry Pi libraries for the sensors, so there was no need in implementing those on my own.

Overall Architecture

Daemon

I have written a daemon that communicates with the sensors and saves data into a file every five seconds. For the greatest portability, the file contains plain text data separated by spaces and new lines (any software can read plain text and analyze it, unlike binary data). After a day of recording, the average file size comes out to be 1M. At midnight, once the daemon finishes the file for the day and continues to a new file, crontab executes another program that fetches the weather data from open-meteo using curl, parses JSON, and produces a chart PDF. This way, each PDF provides both indoor temperature and humidty for the day. At most, this produces 31 API calls to open-meteo per month; the free plan allows up to three hundred thousand calls per month.

Example PDF

Front-end

The front-end was written utilizing HTML, JavaScript, and TailwindCSS. The JavaScript polls the server API every two seconds for new data and updates the charts. Also, the simple charts that display the trend line are drawn onto the canvas with JavaScript.

Back-end

Since I have not written the server, yet, everything relies on nginx and fastCGI module. This way, I can use simple Bourne Shell scripts to fetch data from storage and send off as plain text. Of course, to avoid the headache of constantly hunting for the Pi’s IP, the router needs to be configured to assign a static IP for the Pi.

Convenience

I also display temperature and CO₂ in the status bar of GNOME and MacOS. On Linux, to help comes a GNOME extension that allows executing arbitrary scripts and displaying their outputs in the menu bar called Executor; On MacOS, I have written my own program in Swift utilizing AppKit that provides the same functionality, called CmdBar.

Results

This server has been reliably running and collecting data for almost two years, since March 11, 2024. The longest uninterrupted by electricity outages and accidentals disconnects uptime was over half a year. I am yet to sit down and do any sort of analysis on this data. However, at the time of writing this, I am not sure what use it may provide (compared to, say, monitoring of outside weather) since the indoor values are greatly influenced by the air conditioning, purifiers, etc. But, I guess it can provide some useful insights, like the average indoor CO₂ over the year. If it was worse this year than the previous, it should give a hint to start making changes (i.e. issues with ventilation).

Overall, the project provided a lot of interesting learning experience, touching many areas of programming, like system administration, daemons, UNIX signals, making HTTPS requests with curl, parsing json, writing charts, producing PDFs, API, etc.