The purpose of this assignment was to analyze point patterns. Data used in this presentation was obtained from the City of Vancouver Open Data Catalogue. The data sets included fire halls, disability parking, schools, street lighting panels and sewer manholes. This collection of data sets contained a varied number of points as well as dispersion patterns. Two methods were used, quadrat and average nearest neighbours.
A script tool was created in ArcMap using python and ArcPy. The user inputs a point data set, a study area, quadrat size (user selected or optimal size calculated automatically) and the test significance level. The quadrat analysis generates a fishnet (a square grid) covering the study area and then calculates how many points fall within each grid. The variance between the quadrats is calculated as well as a test statistic value to determine the pattern of the points. The output is a summary of the results and the point pattern type. In addition a fishnet grid is generated to visualize the spacing and the distribution of the points relative to the grid. A very small quadrat size can results in a large number of empty quadrats. As the quadrat size is increased more points will be located in each quadrat and the pattern will disperse with increasing sizes.
Nearest Neighbour Analysis
The average nearest neighbour tool looks at proximity of points. The nearest neighbour index is determined based on the average Euclidean distance between each feature and its nearest neighbour and its compared to the total study area. The output includes the z-score and the p-value associated with the pattern
The pdf poster presentation for this project can be viewed here.