Why is it that to correctly measure pollutant concentrations in the atmosphere, not only quality equipment but also advanced data analysis methods are required?

Manufacturers tend to#nbsp;hide this fact, but any device does not work correctly from time to#nbsp;time. (Yes, it#nbsp;does not feel safe to#nbsp;talk about it#nbsp;because of#nbsp;the nature of#nbsp;competition in#nbsp;new, emerging markets.) At#nbsp;the same time, it#nbsp;often works properly at#nbsp;such moments, but operating conditions go#nbsp;beyond acceptable limits.

An#nbsp;example well known to#nbsp;many sensor manufacturers is#nbsp;that at#nbsp;high levels of#nbsp;humidity, dust particles can absorb water and "stick together", effectively increasing their size and mass. This makes it#nbsp;difficult to#nbsp;correctly measure the dust concentration by#nbsp;optical methods.

One might argue that datasheets contain a#nbsp;description of#nbsp;the equipment’s operating conditions. Yes, indeed, but do#nbsp;you know anyone who always remembers what is#nbsp;said in#nbsp;technical manuals? We#nbsp;hardly. As#nbsp;a#nbsp;result, the user may rely on#nbsp;incorrect data when reporting or#nbsp;making decisions.

A#nbsp;large part of#nbsp;our business is#nbsp;to#nbsp;constantly increase the accuracy of#nbsp;measurement data, and this includes technical aspects (relating to#nbsp;engineering and software development) as#nbsp;well as#nbsp;communication aspects (providing a#nbsp;proper explanation to#nbsp;the user).

The technical challenges include identifying and avoiding cases when the device’s operating conditions are not met. In#nbsp;the above example of#nbsp;dust particles at#nbsp;high humidity levels, the air can be#nbsp;dried before measurement or#nbsp;the particle calculation algorithms can be#nbsp;adjusted.

There are also more complicated cases that can only be#nbsp;detected using an#nbsp;AI-based service that continuously checks the reliability of#nbsp;measurements. The development of#nbsp;such a#nbsp;service is#nbsp;currently one of#nbsp;our key tasks.

For example, our algorithms can already evaluate the reliability of#nbsp;measurements (for a#nbsp;wide range of#nbsp;pollutants) for any specific monitoring site based on#nbsp;the analysis of#nbsp;measurements from all monitors in#nbsp;the network.

Anyway, we#nbsp;find it#nbsp;essential, along with developing data analysis methods that increase the reliability and accuracy of#nbsp;measurements, to#nbsp;openly communicate with our customers about such technological constraints, as#nbsp;they are a#nbsp;part of#nbsp;the natural physics around us.

That is#nbsp;why, in#nbsp;our latest platform updates, we#nbsp;already show periods when measurement conditions could not ensure data reliability and explain why. Paradoxically, this makes our data even more reliable.