One of the most important parts of working with raw sensor, also known as signals, data is to perform some sort of manipulation to remove any sort of noise from the raw signal. Some basic methodologies used include windowed mean, windowed median, mean subtraction (after N samples), etc. These are great to use when verifying the functionality of a sensor, however we usually need more sophisticated approaches to extract better information from our raw signals.
We will be going through a few calibration algorithms to get a better understanding of what is occuring. Moreover, some source code is provided to help you get started.