Yeah, calibration is really the line that separates the hobbyist sector from the much smaller group of people that are trying to use these little open source controllers for research. The frustrating thing about it is the engineers advice is always the same " buy a more expensive sensor and use their calibration". Which is fine if you are doing a standard task, under standard conditions. But most of the researchers are trying to do things that the sensors were not designed for in the first place - For example, most of our temperature sensor IC failures have been because of pressure at depth… which of course they were never designed to withstand. So calibration is always required.
I’ve been developing a new method for reading thermistors with the Input Capture Unit on pin D8. Micro-controllers count time much more precisely than ADC’s measure voltage, so this new approach delivers more resolution than 16-bit oversampling in about 1/10th the time & power