Forum Sensor.Community

Map 2008/50/EC views / f.e. Yearly average value plot for PM2.5

Air quality norms like European Directive 2008/50/EC are often based upon longer averaging periods. Best matching maps.sensor.community view currently is the AQI view.

I would like to see views that show data in relation to the european limits.

PM2.5 yearly average (EU limit 25 µg/m3)
≤ 10 µg/m3 = dark green
10-20 µg/m3 = light green
20-25 µg/m3 = yellow
25-30 µg/m3 = orange

30 µg/m3 = red

PM10 yearly average(EU limit 40 µg/m3)
≤ 20µg/m3 = green
20-60 µg/m3 = orange / amber / yellow

60 µg/m3 = red

90,4 percentile of PM10 daily (36th daily highest mean = allowing 35 exceedances of the 50 μg/m3 threshold over 1 year)
≤ 20 µg/m3 = dark green
20-40 µg/m3 = light green
40-50 µg/m3 = yellow
50-75 µg/m3 = orange

75 µg/m3 = red

There are actually ca. 14.000 active devices, each is sending ~600 values per day.
Generating the file with the 24 h average values needs nearly 4 minutes. Your ‘wishlist’ would need too much resources for our volunteered project.
But you can download all the data at https://archives.sensor.community/ (even pre-packed with a file per sensor type and month).

1 Like

@ricki-z The correct link is singular: https://archive.sensor.community/ and https://archive.sensor.community/csv_per_month/ contains the pre-packed files per sensor type.

These monthly files contain data like:

sensor_id;sensor_type;location;lat;lon;timestamp;P1;durP1;ratioP1;P2;durP2;ratioP2
1368;SDS011;678;49.242;9.334;2021-11-01T00:00:00;0.10;;;0.10;;
32872;SDS011;19224;52.462;16.898;2021-11-01T23:07:55;31.73;;;18.65;;
28668;SDS011;15852;51.44818735;6.56813265;2021-11-01T23:07:55;3.38;;

including invalid PM measurement values like:

13441;SDS011;6793;55.662;13.114;2021-11-01T00:00:00;0.00;;;0.00;;
23220;SDS011;11779;52.680;13.614;2021-11-01T00:00:00;1999.90;;;999.90;;
43574;SDS011;29171;48.196;16.362;2021-11-01T00:00:00;1999.90;;;999.90;;

And most likely invalid data (the change that PM2.5 is exactly the same as PM10 for an average of 5 measurements each 1 second apart is very low), such as:

1368;SDS011;678;49.242;9.334;2021-11-01T00:00:00;0.10;;;0.10;;
38316;SDS011;26181;58.27205456;11.68852172;2021-11-01T00:00:00;0.10;;;0.10;;
12447;SDS011;6288;51.384;6.742;2021-11-01T00:00:00;0.30;;;0.30;;
59655;SDS011;45503;52.81536280;6.82792530;2021-11-01T00:00:00;0.77;;;0.77;;

Some of the measured values raise my eyebrows:

3953;SDS011;1991;48.682;9.030;2021-11-01T21:51:52;1267.73;;;999.90;;
37863;SDS011;23703;47.24256066;8.71710151;2021-11-01T21:51:54;0.17;;;0.00;;
56547;SDS011;42443;57.608;39.862;2021-11-01T21:51:54;1999.90;;;817.83;;

We can’t guarantee that all devices are working correctly (technical problems, sensors at end of live, …). And some users really send data each second. The Sensirion SPS30 is known for the problem that the PM2.5 and PM10 values may be the same after a longer time.
If we find such sensors then we try to contact the owners so that they can solve the specific problem. But this may need some time as we are a volunteered citizen science project.

For now I will try my luck with aqi.eco for my sensor. Aqi.eco seems to be able to output yearly averages and bins daily averages. Though no open source mapping there (closed source Google maps).

1 Like