Humidity correction or disclaimer for lack thereof

[LuchtwachtersDelft]

Several air quality researchers are increasingly signaling that consumer grade sensors and citizen science sensors do not take into account the influence of humidity on particle concentration readings. They fear it leads to citizens and policy makers drawing wrong conclusions. (https://www.researchgate.net/publication/325416608_Suitability_of_the_Low-Cost_SDS011_Particle_Sensor_for_Urban_PM-Monitoring)

Especially above 70% relative humidity the difference between actual PM2.5 and measured PM2.5 concentration grows exponentially. Which is why the operating range of the SDS011 has an upper limit at 70%. And, below 20% relative humidity the sensor actually underestimates the concentration.

While openSenseMap already warns about raw data, we could perhaps clarify the presentation on the map further. For example:

• Describe how the sensors are influenced by relative humidity.
• Provide correction models that can be applied to the raw data.
  Documentation: https://github.com/opendata-stuttgart/meta/wiki/Luftfeuchte-Korrektur
  Example: http://www.ogvt.org/pm24h.php

Besides humidity, sensors also drift over time and there may be variation between devices straight from the factory. For that we should look into calibration.