An averaged value heat map defines colors not by the sum of contributed values, but by their weighted average. This allows you to visualize values where the sum is not meaningful but the distribution is of interest.
This heat map mode is comparable to the concept of a prediction surface in geostatistics. You can use it when the measured phenomenon is assumed to be of a continuous nature (such as signal strength), so that a few sample points can create a continuous value surface.
This mode uses a normalized two-dimensional Gaussian kernel (normal distribution with a standard deviation of 1) with a user-defined
bandwidth applied. Since the normal distribution is of infinite range, the
bandwidth value defines the distance to the first standard deviation.
The weight of an individual or aggregated data sample in any pixel of the rendered heat map is determined by the value of the kernel function at a given distance to the location of the point.