I recently hooked one of these humidity sensors up to a raspberry pi through an ADS1115 16 bit a-to-d converter. I also have a thermistor hooked up to one of the other ADC channels. On successive reads of the sensor, I was getting about +/- 0.5% RH readings. This seemed pretty noisy to me. I tried taking 10 samples and averaging them, and that still didn't help much. I hooked a scope up to the sensor output and saw a lot of spikes in the output trace with a peak-to-peak of about 300mv at a frequency around 1 - 5khz. I thought maybe this could be noise coming from somewhere, but my thermistor channel was much quieter. I also noticed that in the datasheet for the HIH-4030, they show an 80k load resistor hanging off the output. I tried adding this, and on the scope there wasn't much effect. So what I finally did was instead of 10, I took 1,000 samples and averaged them. That made the result appear to be much more stable, maybe +/- .1% RH over a short period, and +/- .3%RH over a few minutes, which I think could just be changes in the actual humidity. Just thought I'd share in case anyone had similar problems.