A Year of Measuring Air Quality

Air pollution seen over Birmingham

It seems hard to believe, but Push Bikes joined the Luftdaten Sensor Community a year ago. Quite a bit has happened since then. For most people the covid pandemic has been the most significant event day-to-day during that year, something that has made the issue of airborne particulates even more significant, as this form of pollution has been found to have a significant effect on one's chances of surviving covid. We've also had the news of a death being officially recorded as being due to particulates pollution.

The original project has grown from its German "Luftdaten" origins into a multinational, multilingual "Sensor Community". It is expanding to cover other forms of environmental pollution. There is a new ambient noise sensor currently undergoing testing in the field, and they have just started work on nitrogen dioxide measurement.

Unfortunately early in the first year of operation, just after the first lockdown, the dust sensor failed. It maxed out to some impossible reading that made a mess of the graphing. Various problems, including lockdown, made recovering the measurement node for repair tricky, but eventually this was done. Since the public graphs were now useless, a bit of Linux shell scripting was required to recover the data from "the cloud" so I could roll my own graphs. To minimise the chance of future data loss, the data is now transferred both to the cloud and a local Raspberry Pi single board computer.

The repaired measurement node was moved to a new home. Before the move the results were most definitely not what was expected, suggesting it would be more interesting to have the node in a very sheltered position instead of next to a busy road. Unfortunately the node's new location was not ideal, as it was too close to a boiler plume, and it caught the morning summer sunshine. It was eventually moved to a permanent home over Christmas, a job that, ironically for an air quality sensor, necessitated drilling through some asbestos sheet (hence my reluctance to move it). However, the new location is very sheltered and will only see direct sunshine late on summer evenings.

I had noticed that the measured temperature was always too high. That didn't change when the measurement node was moved to its current location, and neither did it change when the sensor was replaced just a few days short of the one year warranty because the humidity reading got stuck at 99.99% (this is due to a design problem with the previously recommended Aosong DHT22 sensor, so I replaced it with the currently recommended Bosch BME280 sensor). It turns out that a bird box is very good at being significantly warmer inside than outside, even when all it contains is some very low power electronics rather than a family of birds. So the measurement node has been modified to allow the sensor to dangle in a short section of washing machine waste pipe just below the bird box. Incidentally, the replacement sensor also measures atmospheric pressure.

In the meantime other measurement nodes have come and gone in the Birmingham area. Two nodes that have been present for most of the year are in the Selly Oak area. Another, which alternates between indoors and outdoors, is in Moseley, and one appeared in Sutton Coldfield for a while. It's interesting that these locations match the locations of areas where cycling is strongest in Birmingham, though I only know that one of these sensors was built as a result of the original Push Bikes article. The others are very mysterious. If one of them is yours, I would love to hear from you.

These sensors have allowed me to issue a couple of warnings on the Push Bikes website about poor air quality. They've also represented an interesting educational journey; more on that soon.