Home Automation - an air quality comparison

And now for something completely different!

I had a disagreement with my wife a while ago. See, I've long suspected that our old house - originally built in the 1950s - has an air quality problem. Just a minor one, but still a problem.

I say minor because, according to my measurements, pretty much the only issue is that in two rooms, the HVAC system isn't able to circulate air almost at all (which isn't surprising considering there are neither air inlets nor outlets in those rooms!)

The air circulation design of the 80s was apparently not quite up to the standards of the 2020s.

Fashioning myself as being data-driven, I had already taken the opportunity to buy a bunch of different air quality sensors to better scope the problem. Having gathered the data, I was going to solve the actual issue by buying more gadgets (air exchangers with Home Assistant integration, if you must know).

But then my wife brought up the massive flaw in my study: I have no control group. And to my chagrin, I had to admit she was right.

If I wanted to be data-driven, I needed to not only have data, but also a reference point to compare it to. What good is knowing the levels of radon, VOC, CO2 and other measurements if I don't know what they should be in a residential house?

I could always Google, but that gives me a wide range of numbers, and I don't know how actionable or relevant that is going to be. So I decided to do the next best thing: find a known good house, measure the same things there, and then compare the results. I mean, if I can find a house that is known to have good air quality, and then measure the same things there, I can at least get a rough idea of how bad our house is.

So I set up a small trial to compare the air quality over the same period using similar hardware in a recently built friend's house and ours.

I don't know how my friends tolerate me.

Here are my findings.

Set up

Our house was originally built in the 1950s, then expanded in the 1980s, renovated in the 1990s, 2000s, and 2010s, and then expanded again in the 2020s. At some point in the past it used to be a log house. It now has a pretty simple air exhaust system that is not Home Assistant-integrated (I didn't want to upgrade it just for the sake of integration).

The control is a house built in the late 2010s with a modern heat-exchanging air pump, inlets and outlets properly distributed around the house, a smart HVAC system, and just overall a normal Finnish house without any apparent air quality problems.

Both sensors were located in the kitchen: in our house, the kitchen is also the "real" living room (being the heart of the house and connected to the upstairs lounge); in the control house, kitchen and living room are a joined space.

This unfortunately means I couldn't measure the air quality in the rooms with the worst air circulation in our house, but I wanted to have the measurements from the same place in both houses, so this is the best comparable location I could have.

I do measure the air quality in the rooms with bad circulation as well, but those measurements are not included in this comparison, since they are not comparable to anything.

Measurements

My comparison period was from 4.3.2026 to 10.4.2026. So about 5 weeks of data.

Driven by convenience rather than scientific rigor, admittedly, but it's still better than having generic numbers from the internet.

Radon

Radon is a radioactive gas that is produced when uranium slowly decays. And uranium is found everywhere in the Finnish bedrock - so radon is a common problem in Finnish houses.

You don't want too much of it in your house - but it's difficult to avoid altogether. In our house, it was one of the first things I measured, and due to slightly elevated numbers (100-200 Bq/m³ long-term average), I implemented a pretty ghetto solution: a small computer fan to slowly remove air from our basement, to stop the radon-rich air from seeping up to the rest of the house. It seems to work, but I wanted to see how it compares to a modern house with a proper air circulation system - and actual radon mitigation system in place!

To my surprise, the control house had a higher radon level than our house. Still not high enough to be a problem, but apparently whatever radon mitigation system they have in place isn't working as well as my ghetto solution.

With my computer fan, I have kept the radon levels in our house below 100 Bq/m³ - and during the trial, the levels were below 50 Bq/m³ on average.

VOC

Volatile Organic Compounds are a byproduct of various household activities and materials, such as cleaning products, paints, and furniture.

They typically have a strong smell, and can cause headaches, dizziness, and other health problems if the levels are too high. One of the problems with VOCs is that they are emitted by a wide range of sources, and can be difficult to identify and eliminate. They spike whenever you do something that involves chemicals, such as cooking, cleaning, or even just opening a new piece of furniture.

In both the control house and our house, the VOC levels spiked often, but stayed within reasonable levels (Airthings considers under 250 ppb to be good).

CO2

CO2 is produced by human metabolism or as a byproduct of burning something.

It is a good indicator of how well the air circulates in a house - if the levels are high, it means that the air is not circulating well, and the CO2 is not being removed from the house. High CO2 levels can cause headaches, dizziness, and other health problems, especially in people with respiratory illnesses.

CO2 levels are a decent proxy for the overall air quality in a house, because if the CO2 levels are high, it means that the air is not circulating well, and the other pollutants are also likely to be high.

The control house had typically fairly low CO2 levels, with one spike during the trial, which was a party with dozens of people in the house. But otherwise, the levels were pretty good.

Our house's CO2 levels were spiking higher - and more often - than the control house, which is a clear indication that the air circulation in our house is not as good as in the control house. We do burn more wood in our house, which could contribute. You can also observe a weekend when my family was traveling, and the CO2 levels were very low.

Humidity

Humidity is fairly self-explanatory, right? In Finland, most houses suffer from way too low humidity during the winter (but high humidity isn't quite as large a problem). Low humidity can cause dry skin, respiratory problems (even nosebleeds!), and other health issues. It can also damage wooden furniture and musical instruments.

The control house had a fairly low humidity level, despite having a humidifier turned on for most of the trial. On the bright side, it seems that the air circulation in the control house is good enough to keep the humidity levels low, even with a humidifier turned on. 😅

My house had a fairly good humidity level, but that shouldn't be surprising with poorer air circulation and a humidifier turned on for most of the trial. The humidity levels were surprisingly unstable, though, spiking during cooking, sauna, and other activities that involve water. I guess the air circulation isn't good enough to even out the humidity levels.

Temperature

The recommended indoor temperature is around 20-22 degrees Celsius, but it can vary depending on personal preference and the season. I'd be fine with 19 degrees, but the rest of the family disagrees, so our AC is typically set to 22 degrees during the winter.

The control house's temperature was stable. I suppose they had the heating set to 21 degrees, and the air circulation was good enough to keep the temperature stable at pretty much exactly 21 degrees.

In our house, the temperature was very unstable. But that's by design: Home Assistant controls the heating based on electricity price, time of day, occupancy, weather forecast, solar production forecast and other factors, so the temperature is allowed to fluctuate a lot.

Additionally, the sensor was located in our kitchen extension and near a window, so it experiences even more temperature fluctuations than the rest of the house, going between 15 and about 22 degrees during the trial. It is designed to fall from 8 pm onwards (to ensure cool nights), and then use the last cheap night-time electricity to heat the house up again in the morning (to make it nicer to wake up to).

But still - that's a lot of fluctuation! Not sure if the home automation is tweaking the heating too much at this point.

Air Pressure

Air pressure is a measure of the weight of the air above us. It can be affected by weather conditions, altitude, and other factors. Rapid changes in air pressure can cause headaches, dizziness, and other health problems, especially in people with respiratory illnesses.

I didn't have an outdoor air pressure sensor during the trial (I do now!), which is a bit of a shame - the air pressure difference between outdoors and indoors is what actually matters!

That said, the air pressure is supposed to be somewhat lower indoors than outdoors, and since the houses are nearby, the outdoor air pressure should be pretty much the same for both houses, so we can still compare the indoor air pressure trends between the two houses.

Well... That's pretty boring. It's the same graph, except the control house's air pressure is about 1 mBar lower than our house's air pressure. I suppose it's because of a more powerful air circulation system?

Other measurements

I'll list these here anyway, but they don't have a reference point to compare to, so they are not as interesting.

PM 1

PM 1 particles are tiny particles that are less than 1 micrometer in diameter. They can be produced by a variety of sources, including cooking, smoking, and burning candles. They can also be produced by outdoor sources, such as traffic and industrial pollution.

In a lot of countries, the air quality outdoors is worse than indoors, so the outdoor PM 1 levels are typically higher than the indoor levels. In Finland, however, the outdoor air quality is generally pretty good, so the indoor PM 1 levels can be higher than the outdoor levels, especially if there are sources of PM 1 indoors.

Looks like someone messed up lighting the wood stove a few times during the trial, causing the PM 1 levels to spike. Otherwise, the PM 1 levels were fantastic - as they should.

PM 2.5

PM 2.5 particles are slightly larger than PM 1 particles, but they can still be produced by the same sources as PM 1 particles. They can also be produced by outdoor sources, such as traffic and industrial pollution.

The graph looks the same as the PM 1 graph, which is expected. I have indoor air purifiers that turn on when they notice any spikes in particle levels, so it's not surprising the spikes are rare and short-lived.

Conclusion

Overall, the air quality was actually pretty comparable. I'll find it difficult to justify buying more gadgets just based on this data.

To capture more, I think I'll need to buy a few more sensors and organize a more rigorous trial, with measurements from more locations in the house, and a longer trial period. Maybe even a third house as a control group, to get a better reference point. And perhaps even upgrade my Home Assistant setup with a more powerful host machine, just to be future-proof. Wouldn't hurt to buy a better outdoor weather station as well, to get better outdoor air quality data to compare to.

But for now, I think I'll just enjoy the data I have and not worry too much about it. After all, the air quality in our house is pretty good, and we are not experiencing any health problems related to it.