To capture patterns in urban form, we have to analyze morphometric characters within a spatial context, taking into the account neighbouring elements. A simple example could be a mean height within 100 metres (measured as a location-based character for each building).
momepy is using
libpysal spatial weights to capture the relationship between elements, in this case, buildings, in the form of a binary matrix (1 = neighbours, 0 = not neighbours). Spatial weights are an essential part of
momepy and are used on many occasions and in various forms. However, they should always be based on the unique ID of element, not index (there are a few exceptions documented in API).
This section covers: