Distribution of temperatures

Now that all of or cameras have been upgraded to report their temperatures to the cloud we have the opportunity to examine the data collected.

The first thing we looked at was how the average temperature reported by our cameras compared to the actual average. This Harvard study reported that the average human temperature is 97.5°F (36.4°C). The average temperature across ~1500 readings from our cameras was 35.8°C. 

This lower average is expected because we are estimating body temperature from a measurement of forehead skin temperature and we know that if you are exposed to cold air you will get a lower reading. We are planning to do some research to see if we can adjust for this. This could also be due to one or more cameras not being calibrated properly.

The next thing we looked at was a histogram of the recorded temperatures. This showed the expected distribution of temperatures, with a bunch of lower outliers as expected.

Body temperature varies through the day by about 0.5°C. We tried to see this by plotting our data vs the time of day. However, as you can see there is no obvious diurnal variation. This requires a little more examination and maybe it will come out when we are logging the temperatures against individuals.

Another interesting thing to look at is how the averages vary across cameras. The graph below shows that there is some variation and that those that are reading significantly different than average should maybe be re-calibrated. However it could also be due to other causes, such as being in a colder part of the country. It may be useful to be able to see how your camera(s) compare to these averages.

These are just a few of the statistics we were able to pull out from a quick analysis. Please let us know if there is anything else you think we should look at.