Device Validation

We have collected a lot of data now and I think we are becoming comfortable that devices are providing reliable readings. Do you agree?

We have not though carried out a formal validation in of devices in the field and I think we should consider doing this. A validation exercise would need 2 or 3 co-located devices close to a known (calibrated) sensor and would need to run long enough to experience a range of environmental conditions (especially humidity).

Others have run this kind of study so we can lean on their conclusions but wouldn’t it be better to run our own? The difficulty is we only have one calibrated device and that is Freetown Way (Hull). I wonder if we could get permission to co-locate some devices there?

Please voice your thoughts.

1 Like

I thought about this today when I noticed the identical peak seen on PM2.5 between 1400 and 1500 at https://graphs.connectedhumber.org

I think the occasional sensor seems to read low, but I do wonder if that relates to housing or placement. It would be interesting to co-locate a couple of sensors in identical housings at the FTW site and monitor over a period of time, but it’s interesting even without doing this how close the sensors readings are across those deployed.

Another thing to think about is how accurate do the sensors need to be? For me it’s probably enough to say AQ is low, medium, high or very high.

1 Like

The overall pattern of our devices is very similar but there are occasional peaks on some and not others. That level of detail could be where the interesting stuff is lurking.

1 Like

Use the Sheffield sensor as well, it’s co-located already, one tick.

Here’s the link to it.

1 Like