Another update:
High pollen counts (source) don’t appear to affect any of the measured values, incl PM1/2.5/10.
NO2 values continue to be erratic. Keeping the phone near the device so it can sync frequently is supposed to help with that, and it does indeed appear to smooth out crazy oscillations, but values still go implausibly high.
VOC appear to go way up when leaving the car parked in the sun. The car is a few years old now, but maybe it needs some more baking out 
I once had a cheap air quality tester. VOC would spike greatly in garbage and certain places in garage. Overall indoors was slightly worse than outdoos.
I’ve been noticing the same behaviour, despite frequent syncing. It’s especially noticeable now during the lockdown here in Paris: Despite there being virtually zero cars I register values in the range of 100-250 Flow AQI, which seems highly implausible to me. Interestingly I notice that this only seems to happen once starting to move (i.e. indoors the values are low & smooth, but as soon as going out the values are all off the chart).
That gave me the idea to take the device to a gas station, and indeed there is was small VOC bump! Not surprising, but could compare against other gas stations that are either busier, or that have older pumps that are less-well sealed, to see how much of a difference it makes…
Just set up a fan to blow air over my Flow 2, but that didn’t seem to confuse it
According to this source, NO2 sensors can be thrown off by ozone and chlorine (unlikely in our cases), and are sensitive to changes in temperature/humidity/pressure. Controlling for these factors may be trickier when moving around outside…
Oh yeah, that makes sense. I’m sure all of those parameters are thrown off significantly when moving from my desk to outside!
Ever so often, I see large PM10 spikes, but then when I look again a few minutes later, Plume appears to have decided that it overreacted…
Kept the windows in the bedroom closed last night. Netatmo shows CO₂ creeping up throughout the night, until I open the window in the morning. Subjective air quality was poor, but not headache-inducing bad.
Meanwhile, the Flow maxes out at an AQI of 13 for VOC (anything less than 20 is considered “good”).
So while changes to VOC levels do roughly mirror CO₂, VOC levels alone aren’t sufficient to help me decide on how much ventilation a room needs.
I got a Netatmo last week and look forward to dive into that data. In the meanwhile I got around to looking into my Flow data a bit more. As it at least feels like there’s a lot less car traffic in Paris these days I thought I’d look into whether I could find any changes in the data relating to the pre- & post-lockdown time that I have sampled.
Being a bit lazy I just calculated daily average values for the 4 categories that the version 1 of the Flow can track (NO2, VOC, PM2.5, PM10) and compared the collection of averages that I got from that.
I’m sure one could do more detailed things, e.g. comparing location & time-of-day/weekday matched samples, but given that my own daily routine hasn’t changed too much between before/after I wouldn’t expect there to be too large systematic differences.
What I found was that the NO2 and PM10 levels show a decrease post-lockdown (which also comes out statistically significant when doing an overly simple Mann-Whitney-Wilcoxon test), while the VOC & PM2.5 don’t show any decreases.
Either way, it’s not a huge difference in the mean values and as the post-confinement period falls into summer time – with Parisians going on holidays – I’ll have to collect more data 
Finally got an “opportunity” to track some indoor and outdoor pollution from wildfires…
Here’s yesterday’s air quality data from my Flow:
08:30 bedroom, opened window for 15min (official AQI 180)
09:45 home office
10:45 living room
15:00 went shopping (official AQI 206)
15:45 living room
19:30 kitchen, pan-seared fish
11:15 bedroom, opened window for 15min (official AQI 175)
I have a small air purifier that I kept running all day, and that following me around the apartment (except into the kitchen).
What did I learn?
I don’t know/care if the PM2.5 values reported by the Flow are accurate, but they appear plausible, and are useful for tracking exposure to smoke. PM1 values are almost identical (but have gaps). PM10 tends to go up with PM2.5, but also reflects other (?) factors. Not sure what NO2 is supposed to do, and VOC continues to be a poor substitute for CO2.
When opening a window, pollution gets in much faster than stale air gets out.
Running my air purifier at full blast, it takes about 45min to clear out a room after opening the windows.
Even with all windows closed, some pollution creeps in, but can be kept at bay by running the air purifier on low.
Pan-searing food isn’t the best idea when the windows can’t be opened…
Very interesting results. They seem coherent with what I’ve observed here running an air purifier with a visual display of air quality in our main room. With doors and windows closed the air quality slowly improves, but even a quick opening of the door will drop the air quality for up to an hour.
Should mention that my air purifier has an activated carbon pre-filter, which should adsorb some VOCs. This might contribute to the discrepancy between the VOC and CO2 measures.
Just had the dishwasher replaced, and noticed that the new dishwasher doesn’t “leak” nearly as much PM10 as the old one did:
That’s fascinating - never would have occurred to me that a dishwasher was a source to watch out for!
According to this source, many of the VOCs found in cars are carcinogenic (or otherwise problematic), so airing out the car a bit before driving off is probably a good idea!
Has anyone here used a Sensio AIR? I’m curious how well our air purifier setup handles pollen; this isn’t something the Flow appears to be capable of measuring, and this device is slightly outside the budget 
I’m using an Atmotube Pro which seems to have some validation. In a future i’ll try to find associations between sleep quality and quantity vs nightly air quality (temp, humidity, air pressure, particles).
Sensors on a Plane!
Originally shared at the QS Seattle meetup last month.
Took my Flow along on a long flight (LX1068, SEA-FRA, A330), followed by a short flight (LH490, FRA-ZRH, A320), with a few hours at an airport lounge in-between. Didn’t mark the exact times I got on and off planes, but it’s obvious enough:
The same trip in reverse (minus the lounge):
Notes:
Here’s another long flight from a different trip (FI680, SEA-KEV, Boeing 737 MAX 9), followed by a shorter flight (FI568, KEV-ZRH, Boeing 737 MAX 8):
Notes:
The VOC sensor died before the return flight, so no more data 
Dishwashers that were worse than outdoor pollution is this thread right?
