Using spatial weights matrix

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:

• Generating spatial weights
  • Queen contiguity
  • Distance
• Examples of usage
  • First order contiguity
  • Higher order / distance
• Measuring diversity
  • Queen contiguity
  • Range
  • Theil index
  • Simpson’s diversity index
• Using two spatial weights matrices
  • Building adjacency
  • Mean interbuilding distance