Pattern Analysis

I used ArcGIS Pro 3.0.1 and 2.9.3 to solve this problem. The main geoprocessing tools that I used include “Average Nearest Neighbor”, “High/Low Clustering”, “Multi-Distance Spatial Cluster Analysis”, and “Spatial Autocorrelation”. There are MXD files provided by the class, which are maps of the calls for services from Battalion 2, Fort Worth, TX, and library patrons from the Oleander Library.

(1) For exercise 1, I imported the map layer into ArcGIS Pro and only showed false alarm data using a definition query. Then I used the Average Nearest Neighbor to identify if the false alarm call locations are clustering. This step is only using the geographic information of the false alarm call. Then I examined the z-score and p-value to determine if I could reject the null hypothesis that false alarm calls are randomly distributed.

(2) For exercise 2, I imported the map layer into ArcGIS Pro. Then I used the High/Low Clustering tool to identify if clustering occurs for the locations of calls for services, specifically, if high-priority calls are clustering. Then I redo the analysis using different distances and compared the G index and Z scores. Then I was able to identify if clustering occurs and the distance with the most prominent clustering.

(3) For exercise 3, I imported the map layer and used the Multi-Distance Spatial Cluster Analysis tool that includes all neighboring features in calculating Ripley’s K. I also included 99 permutations and compared a range of distances. Then I calculated the difference between the observed K index and the upper limit of the confidence envelope. Then I made a graph that shows the changes in difference over expected K (the peaks suggest the distance in which most clustering occurs).

(4) For exercise 4, I first imported the new map layer of library patrons. Then I did a spatial join with a 300-ft grid and the patron locations. Then I used the definition query to only show grids with more than 0 joined counts. Then I run the Spatial Autocorrelation tool with a range of distance (a tool that examines both attributes and location). Then I compared the z-scores and confidence level at each distance and identified the distance with the highest z-score. All exercises resulted in maps that showed the results.

Identifying spatial patterns is a very useful skill to have. Knowing whether a feature is clustered or not could have implications on management decisions. I also appreciate the mention of how to create a grid from the tutorial (Create Fishnet). Before learning GIS, I always wondered how to create a grid that could be used for determining sampling locations. A possible scenario could be that a Citizen Science specialist needs professional help in analyzing the spatial clustering pattern of citizen science participation.

Problem description: As a GIS specialist, I am asked to perform global spatial pattern analysis to determine if participants of eBird (bird reporting citizen science project) report checklists that are spatially clustered in North Carolina (e.g., checklists clustered in higher income/white neighborhoods).

Data needed: Tabular data of eBird checklists in NC.

Analysis procedures: I would import a point layer of eBird checklist locations in NC. Then I will use the Average Nearest Neighbor tool to identify if clustering occurs by examining the z-score and p-value of the analysis. I could also use the Multi-Distance Spatial Cluster Analysis tool to identify the distance where the most prominent clustering occurs.