Most often, the term humidity is actually relative humidity in percentage (%) rather absolute humidity. The latter refers to the vapor concentration as a ratio of mass to volume (grams/m3). The relative humidity (RH, %) is the ratio of the mass of vapor at the point of measuring to the mass of vapor at saturation of vapor tempeture (100%). Another term is called dew point (unit: ℃) indicating the temperature that the vapor in that sampling air condensed. Other terms related to the humidity are here.

According to the graph below, the dew point depends on the original relative humidity (how much vapor that air contains) at which temperature. Let say, a sample air has a relative humidity of 70% at 30 ℃, using this graph, the dew point of this sample air is around 24 ℃, but the that air only have 20% RH, then the dew point is ~ 4℃.

Source: Wikipedia

For my case, the graph below shows the mirroring lines between the temperature and the relative humidity but this observation ignores the underlining facts that the relative humidity is calculated based on the temperature. So with a same amount of absolute vapor in a sample air, a higher temperature results in a lower relative humidity (because of that tempreature, that air can hold more vapors relatively the absolute capacity of holding vapors at the latter temperature).

The next question is what are the choice for a humidity sensor? There are measuring principles based on change of resistance, capacitance, and weight. Many reviews online offers better explaination, so you can browse like one in here.

As of 2018, the humdity sensor for a hobbist are mostly resistant or capatative types. An excellent comparison posted by Robert Kandrsmith is here. The tested sensors included AM2302, SHT71, HTU21D, Si7021 (used the same driver as STH21), BME280. The accurate range is ∓3% in the range of 10-80 %RH. The weather in Hanoi could reach to a typical 80-90% and the recorded values in the range bears more uncertainty. My weather station used two version one is Si7021 but the sensor is marked at SHT21 and the other is DHT22. Adding HTU21D is not possible since the I2C address of Si7021 and DHT21D is the same (0x40), and no option to change the address is provided. (ussually through copper trail and solder blob)

Applications of the relative humidity addressed in two main areas: labratory and human comfortability. In gravimetric analysis or analyzing of weight change especially with a sample contains water, the relative humdity has to be specified so that the weight of weater vapor in the sample is taken into account.

The relative humidity can weight on human confortablity. A wetted (or high (relative) humidity) plus a high temperature with a low connection (such as no wind/fancing) reduces the evaporation of water that our body employed to cool off after excercising or have extra heat than its normal state (36.5℃). So we execret sweat instead of vapor. So the feeling being covered by sweat and cannot get rid of it because the air around is almost saturated (I'm full, say no to more vapor, said the air).

A dried air is not helpful either. Dried air can lead to crack skin, running nose, bleeding in the respiratory track because of the imbalance of the vapor in the air. In this case, the air sucks the water out of your skin. Did you recalled running in cold, dried weather and the nose was sketchy and you smelled funny.

But not to worry about the comparison, engineers proposed a heat index as the combo term to take temperature and relative humidity into a figure number to indicte (mostly) hotness (of the air). According to Wikipedia, at 24℃ and 0%RH, the air temperature feels like 21℃, but at 24℃ and 100% RH, the air feels like 27℃.

I hope this writing could give some info to start up with relative humidity. Thanks.