[throwaway9d0291]

The study is interesting but I wouldn't say it "did rather well", as they applied a calibration to it first:

"A regression analysis was performed to improve the accuracy of the Foobot FBT0002100 data relative to the GrayWolf data. Field calibration equations were then produced from the calibration dataset using the results from the GrayWolf instruments as dependent variables and the Foobot FBT0002100 as independent variables and tested on the validation dataset. An analysis in SPSS of the linear, quadratic, and cubic models was performed individually for each parameter to find the most accurate equation."

You can't do that unless you happen to have access to a different research quality sensor. If the Foobot had that out of the factory it'd be nice though.

[magicalhippo]

That's not how I read it.

"Field calibration equations were then produced from the calibration dataset [...]" (emphasis mine).

Table 3 shows the calibration dataset for tVOC, which at least to me looks rather ok except for the outlier E.