Spatial statistics analysis
Cluster and oultier analysis
The final TOD maps are not completely useful for planning applications. It is necessary to identify regions of high and low TOD values so that planning decisions can be made accordingly. While this could be done via visual inspection of the maps, this can be a subjective process.
So, to remove any subjectivity and identify these regions of high and low values, a spatial statistics analysis must be done.
The analysis that best meets the needs is a cluster and outlier analysis (also known as Anselin Local Moran's I). This analysis identifies clusters of high and low values and measures the degree of similarity or dissimilarity of neighbouring features.
Outputs
The output code is to be interpreted as follows:
HH = a feature with a high value that has neighbours with high values
LL = a feature with a low value that has neighbours with low values
LH = a feature with a low value that has neighbours with high values
HL = a feature with a high value that has neighbours with low values
Conceptualising the spatial relationship
The spatial relationship between features is important in spatial statistics. There are two parameters that have to be defined:
-
The relationship between distance and influence, and
-
The distance threshold.
The chosen relationship was 'zone of indifference', and the threshold distance was 282.5m. For more details on how this spatial relationship was chosen, refer to the full document.
This function describes the relationship between the distance and level of influence for a particular target feature and its neighbouring features.
The threshold distance of 282.5m means that the adjacent neighbouring features have the same level of influence as a given target feature.
Click image to enlarge
Click image to enlarge