Topic 10 Flashcards
(18 cards)
pattern analysis
fundamental form of spatial analysis
employs both spatial processing and descriptive statistics
study of spatial arrangments of point or polygon features
point pattern analysis
techniques that examine the proximity and density of points in space
Ripleys K function
multi distances (how clustering changes with changes in neighbourhoods)
what is teh pattern of a point pattern called
point dispersion
what is the point pattern analysis
a simple metric of point dispersion to identify whether a point pattern is regular or clumped
based on neighrest neighbor
point pattern analysis interpretation
if variance = mean, pattern is random
if variance exceeds the mean, pattern is clustered
if variance is less than mean, pattern is regular
pattern analysis on polygons
Getis-ord general G
high/low cluster analysis is another common technique
useful for detecting clusters of polygons with high or low attribute values
spatial autocorrelation
metrics assess the relationship between proximity and similarity of the attribute of a feature
essentially a quantitative expression of toblers law
positive spatial autocorrelation
features that are close in location are also similar in attributes
negative spatial autocorrelation
features that are close in location tend to be dissimilar in attributes
zero spatial autocorrelation
attributes are independent of location
the weights matrix
defines the locational similarity of the i (column index) and the j (the row index)
1=adjacent
0= not adjacent
measuring spatial autocorrelation
any measure is simply an index that combines some measure of locational similarity with a corresponding measure of attribute similarity
morans I
widely used measure of global spatial autocorrelation,
ranges from -1 to +1
distance effects and spatial correlation
indices (like morans I) are measure of spatial association among attributes and provides insight into the distance affects
hot spot analysis
Getis-Ord gi*
Local Moran’s I
Getis-Ord Gi*
looks for hot and cold spots
evaluates z-scores to provide statistical evaluation of the significance of the high or low values
Local moran’s I
cluster outlier analysis
evaluates if a point is significant