This paper presents a comprehensive discussion of the deployment, calibration and performance of a dense array of low-cost sensors. The paper is interesting, shows that the sensors have potential for scientific application and is clearly written. The initial analyses explaining differences in local concentration patterns was very insightful and interesting. I enjoyed reading this paper. I recommend publication after clarification of some minor issues.

I found the discussion of humidity slightly confusing. Humidity has at least two effects. First water contributes to the total pressure requiring a correction if one is expressing a dry air mixing ratio. Second water affects the spectroscopy of CO2 in the NDIR sensor both through pressure broadening and through spectral interference. The authors begin by saying they make a correction to dry air. I was confused about when in the sequence of their analysis they make the correction to dry air and whether it makes any difference if they do that at the beginning or end of their analysis. I'm not challenging what the authors are doing, just hoping for some clearer procedural description. Second, is more subtle. A comparison of calibrated data to a reference should be flat if plotted against RH, absolute humidity or temperature. It's unclear to me if the authors checked their calibration in this way. A brief comment on the residuals of the low cost sensors vs. the reference along these coordinates would be helpful in evaluation of the data set.