Problem:
The goal of this project is to evaluate 4 different spatial pattern analysis techniques to determine the chance patterns are occurring due to random chance.
Analysis Procedures:
ArcGIS was used for this analysis. Tools included average nearest neighbor, calculate distance band from neighbor count, multi-distance spatial cluster analysis, and spatial autocorrelation. The data was provided by the NCSU GIS 520 instructor, it was from the GIS Chapter 8 Tutorial.
The assignment consisted of four different exercises to evaluate tools that allow pattern analysis in spatial data. For exercise 1, emergency false alarms from February 2015 were explored to determine if there is “clusters” or specific areas that have false alarms. This was done using the average nearest neighbor tool. Next, exercise 2 was to determine if there was clustering in calls of service to the fire department from January 2015. This was done with a high/low clustering calculations using the calculate distance band from neighbor count tool. This was repeated for distances ranging from 700 to 1200 to determine the highest z-score meaning the distance with the most clustering. Next, exercise 3 explored the calls of service for January 2015 using a different technique. This analysis was performed with the multi-distance spatial cluster analysis where a graph of clustered versus dispersed was created to evaluate the clustering of calls. Finally exercise 4 involved spatial autocorrelation of patron location with respect to the oleander library. This was done using a 300 foot grid. Distances of 2800 to 3800 feet was explored to determine the distance that resulted in the highest z-value and determine if this z-value could reject the null hypothesis.
Results:
Exercise 1 Results 
Exercise 2 Results
Exercise 3 Results 
Exercise 4 Results
Application and Reflection:
The assignment has taught me various ways to determine if clustering is occurring in my spatial data. Understand z-values and their associated p-values will allow me to determine clusters in any set of data. An example would be car accidents within Wake County. If there was a shapefile available within in Wake County for all car accidents than a grid could be created for Wake County and the spatial autocorrelation could be used to determine if there are clusters in the data. These results would allow law enforcement agencies to plan where to patrol more to anticipate and hopefully prevent accidents.
Spatial Problem Solving Portfolio 