Indoor air quality is one of those topics where most people have opinions but few have data. We decided to change that for ourselves by deploying five calibrated air quality sensors across three rooms in our test home and logging continuous readings for 90 days. We tracked PM2.5 (fine particulate matter), VOCs (volatile organic compounds), CO2, temperature, and humidity. The results surprised us in ways that reshaped how we think about air purifiers, ventilation, and the common wisdom about indoor air quality.
The Setup
We used three PurpleAir PA-II sensors (PM2.5, temperature, humidity) and two Awair Element monitors (PM2.5, VOCs, CO2, temperature, humidity). One PurpleAir sensor was placed outdoors for reference. The remaining four sensors were placed in a kitchen, a living room, a bedroom, and a home office. All sensors were calibrated against a TSI DustTrak reference instrument at the start and end of the monitoring period. Readings were logged at five-minute intervals, producing approximately 78,000 data points per sensor over the 90-day period.
We did not change our normal household routines during the monitoring period. We cooked normally, cleaned normally, opened windows when we felt like it, and ran our existing HVAC system on its normal schedule. The point was to capture real-world conditions, not laboratory-clean scenarios.
Finding 1: Cooking Is the Biggest PM2.5 Source by Far
The kitchen sensor showed PM2.5 spikes that dwarfed every other indoor source. Gas stovetop cooking produced average peaks of 150 to 280 micrograms per cubic meter — compared to a baseline of 8 to 15 micrograms per cubic meter. For context, the EPA's 24-hour PM2.5 standard is 35 micrograms per cubic meter, and the WHO's annual guideline is 5 micrograms per cubic meter. A single stir-fry session pushed our kitchen to PM2.5 levels that would trigger an air quality advisory if they occurred outdoors.
The kitchen range hood — a standard recirculating model, not ducted to the exterior — reduced peak PM2.5 by approximately 25%. That sounds helpful until you realize it still left cooking PM2.5 at 110 to 210 micrograms per cubic meter — four to seven times the EPA's 24-hour standard. A ducted range hood (venting directly outside) would eliminate the vast majority of cooking-generated particulates, but our test home, like many homes, was not equipped with one.
Opening a window within ten feet of the stove reduced cooking PM2.5 by 40 to 55% compared to the recirculating hood alone, and the two combined reduced peaks by roughly 60%. This was the single most effective air quality intervention we identified during the entire 90-day study — and it costs nothing.
Finding 2: Sleeping With the Door Closed Traps CO2
The bedroom sensor revealed a pattern we had not anticipated. With the bedroom door closed overnight, CO2 levels rose from a baseline of 450 ppm (parts per million) to 1,200 to 1,800 ppm by morning. Two adults in a closed 150-square-foot room produce enough CO2 to reach these levels within three to four hours. Research published in Indoor Air (2016) found that CO2 levels above 1,000 ppm are associated with measurable declines in cognitive function and sleep quality — and our bedroom consistently exceeded that threshold by midnight.
Opening the bedroom door dropped the overnight CO2 peak from an average of 1,500 ppm to 650 ppm. Cracking a window achieved a similar reduction. Either intervention was sufficient; both together were unnecessary. This finding is consistent with the growing body of research suggesting that bedroom ventilation is one of the most overlooked factors in sleep quality — and it requires no equipment, no filters, and no electricity.
Finding 3: Cleaning Products Are a Major VOC Source
We expected cooking to dominate our VOC readings. It did spike VOC levels, but the highest sustained VOC readings came from cleaning products. Spraying a standard bathroom cleaner in the shower produced VOC levels that remained elevated for 45 to 90 minutes after cleaning was complete. Using a scented floor cleaner in the living room kept VOCs above baseline for over two hours.
The most effective mitigation was simple: open windows during and for 30 minutes after cleaning. Switching to unscented, low-VOC cleaning products reduced peak VOC levels by approximately 60% without any ventilation changes. The combination of low-VOC products and open windows brought cleaning-related VOC exposure to negligible levels.
Finding 4: Air Purifiers Help — But Less Than Ventilation
We ran a HEPA air purifier in the living room during the second half of our monitoring period. It reduced ambient PM2.5 by approximately 30 to 40% and had no measurable effect on CO2 or VOCs (HEPA filters capture particles, not gases). This is consistent with manufacturer claims and independent testing — HEPA purifiers work for particulate matter.
However, simply opening two windows on opposite sides of the house (cross-ventilation) reduced PM2.5 by 20 to 50% (depending on outdoor air quality), CO2 by 35 to 60%, and VOCs by 40 to 70% simultaneously. The air purifier addressed one pollutant category; ventilation addressed all three. In a climate where outdoor air quality is consistently good, ventilation is the more comprehensive and cost-effective intervention. In areas with poor outdoor air quality (wildfire smoke, industrial pollution, high pollen), a HEPA purifier is essential because opening windows would make indoor air worse.
What Changed for Us
After 90 days of data, we made four permanent changes to our household routines. First, we open the kitchen window during any stovetop cooking — this single habit reduced our highest daily PM2.5 exposures by more than half. Second, we sleep with the bedroom door open, which keeps overnight CO2 below 700 ppm. Third, we switched to unscented cleaning products and ventilate during cleaning. Fourth, we run the HEPA purifier only during wildfire season and high-pollen days, rather than 24/7.
The most important lesson from this experiment is that indoor air quality is not a product problem — it is a behavior problem. The interventions that made the biggest differences were free (opening windows, opening doors, timing activities) while the interventions that cost money (air purifiers, upgraded range hoods) made smaller contributions. Data changed our assumptions, and better assumptions led to better air without spending a dollar more than we already had.
Sensor Accuracy and Calibration
Consumer air quality sensors are not laboratory instruments, and setting appropriate expectations for their accuracy is important for interpreting results correctly. The PurpleAir PA-II sensors we used cost approximately $250 each and measure PM2.5 using a laser-scattering method that correlates well with reference instruments under most conditions. Our pre- and post-study calibration checks showed agreement within plus or minus 15 percent of the TSI DustTrak reference for PM2.5 readings between 10 and 200 micrograms per cubic meter. Below 10 micrograms per cubic meter, the relative error increased — a reading of 5 micrograms per cubic meter could actually be anywhere from 3 to 8 micrograms per cubic meter.
The Awair Element monitors ($149 each) measure additional parameters but with lower precision than dedicated single-parameter instruments. Their CO2 readings agreed with our calibrated Vaisala reference within plus or minus 75 ppm — adequate for detecting trends and identifying problem conditions but not sufficient for precise concentration measurements. Their VOC sensor uses a total VOC (TVOC) approach that cannot distinguish between harmful and benign volatile compounds. A reading of "high VOC" could mean harmful cleaning product fumes or harmless cooking aromas. We used the VOC readings as screening tools — elevated TVOC triggered further investigation, not automatic concern.
For homeowners considering their own monitoring project, one well-placed sensor is more valuable than no sensors. Place it in the room where you spend the most time (usually a bedroom or living room) and establish your baseline readings over two to four weeks before drawing any conclusions. The absolute numbers matter less than changes from your personal baseline — a reading that jumps from your normal 10 micrograms per cubic meter to 60 micrograms per cubic meter is significant regardless of whether the sensor's absolute accuracy is perfect.
Sensor types and what each measures
PM2.5 sensors (particulate matter, 2.5 micrometers and smaller): The single most valuable air quality measurement for health. PM2.5 particles are small enough to penetrate deep into the lungs and enter the bloodstream, contributing to respiratory and cardiovascular disease. Sources include wildfire smoke, cooking (especially frying and grilling), candles, incense, tobacco smoke, and outdoor pollution infiltrating the home. Consumer PM2.5 sensors ($30 to $200) use laser light scattering — a laser illuminates particles drawn through the sensor, and a photodetector counts the light scattered by each particle. Accuracy varies: the best consumer sensors (PurpleAir, IQAir) correlate within 10 to 20 percent of reference-grade instruments; the worst (generic AliExpress sensors) can be off by 50+ percent. For the purpose of DIY monitoring, even moderate accuracy provides actionable information — you do not need laboratory precision to know that cooking without ventilation raised PM2.5 from 5 to 150 µg/m³.
CO2 sensors (carbon dioxide): CO2 concentration is the most direct proxy for indoor ventilation adequacy. Outdoor air contains approximately 420 ppm CO2. In a well-ventilated occupied room, CO2 stays below 800 ppm. In a poorly ventilated room with occupants, CO2 can exceed 2,000 ppm within 1 to 2 hours. Above 1,000 ppm, cognitive performance measurably declines — studies show 15 percent decrements in decision-making performance at 1,000 ppm and 50 percent decrements at 2,500 ppm. CO2 sensors ($50 to $150 for NDIR sensors, which are the accurate technology; avoid the cheaper eCO2 sensors that estimate CO2 from VOC readings and are unreliable) provide a simple, actionable metric: if CO2 exceeds 1,000 ppm, open a window or increase mechanical ventilation.
VOC sensors (volatile organic compounds): VOCs are gaseous pollutants emitted by paint, adhesives, furniture (formaldehyde from pressed wood), cleaning products, air fresheners, and cooking. Consumer VOC sensors ($20 to $50, typically included in multi-sensor air quality monitors) detect total VOC concentration but do not identify specific compounds. The readings are useful for identifying periods of elevated VOC exposure (during and after cleaning, after painting, when new furniture is off-gassing) but cannot distinguish harmless VOCs from harmful ones. For identifying specific VOCs (formaldehyde, benzene), dedicated single-compound sensors ($30 to $100 each) provide more actionable information.
Long-Term Implications
The 90-day monitoring period changed our understanding of indoor air quality from an abstract concept to a measurable, manageable reality. The biggest surprise was how much our own behavior — cooking methods, cleaning products, door and window management — determined our exposure levels. External factors like outdoor air quality, traffic, and industrial emissions were measurably present in our data but contributed less to total indoor exposure than our own activities.
This finding is both empowering and sobering. It means that individual behavior changes can dramatically improve indoor air quality without expensive equipment or building modifications. But it also means that most indoor air quality problems are self-inflicted and persistent — they recur with every cooking session, every cleaning cycle, and every night spent in a closed bedroom. The monitoring gave us the data to make better decisions. The decisions themselves require ongoing discipline. The sensors can tell you the room needs ventilation, but you still have to open the window.
The financial case for monitoring is compelling when you consider what it prevents. A single visit from a mold remediation specialist costs \ to \ for assessment alone, with remediation running \,000 to \,000 depending on severity. The respiratory health costs of chronic poor indoor air quality — increased allergy medication, asthma management, sick days — are harder to quantify but equally real. A \ to \ air quality monitor that identifies problems before they escalate pays for itself with the first problem it catches.
The data also provides peace of mind that subjective perception cannot. On multiple occasions during our study, we were convinced the air quality was poor — after cooking a smoky meal, during a humid spell, after using cleaning products — only to find that the measured levels had already returned to baseline. Conversely, there were times when the air felt perfectly normal but the sensors showed elevated CO2 from hours of closed-room work or PM2.5 drifting in from a neighbor's outdoor fire. Perception is unreliable for invisible pollutants. Measurement is not. And the gap between what you think your air quality is and what it actually is closes permanently once you start tracking it.
