Pattern Analysis

Exploring Spatial Patterns in Distribution of Emergency Calls and Library Patron Locations

Problem and Objective

Excercise 1: The Fort Worth Fire Department is interested in knowing if there is any clustering of false alarms.

Exercise 2: The Fort Worth Fire Department is interested in knowing if high priority calls for January of 2015 have any clustering.

Exercise 3: The Fort Worth Fire Department wants to know if high priority calls for January of 2015 exhibit any clustering when also taking into account the relationships of each feature in the specific study area.

Exercise 4: The Oleander Library is curious to know if their patrons are randomly distributed among city blocks, or if they experience clustering.

Analysis Procedures

To find these solutions, I used the ArcGIS Pro software. Specifically, I used the Average Nearest Neighbor tool, the High/Low Clustering (GetisOrd General G) tool, the Multi-Distance Spatial Cluster Analysis (Ripley’s K Function) tool, the Create Graph tool, the Spatial Join tool, and the Spatial AutoCorrelation (Global Moran’s I) tool. The data that I used was from the Fort Worth Fire Department on where the false alarms are located and where all calls for service came from. I also used Battalion boundary data and data from Oleander Library on where their patrons are located.

Once all of the data from the Fort Worth Fire Department has been imported to ArcGIS Pro, I was able to start the analysis. I first used the Average Nearest Neighbor tool with the false alarm data. From this I received a negative z-score and was able to determine that the false alarms are randomly distributed. I then used the High/Low Clustering tool witht the data on all the calls for service in January 2015. This tool helped me determine the calls during this month had the highest amount of clustering at a 900 foot distance band. I then used this same data on calls for service in January 2015 in the Multi-Distance Spatial Cluster Analysis tool. This tool is similar to the High/Low Clustering tool, but it is more accurate, because it takes into account the specific study area and relationship between features. Using this tool, I received the same solution as the High/Low Clustering tool, so this was a good check to make sure the high priority calls are clustering at a 900 foot distance band.

Switching gears, I uploaded the data from the Oleander Library to ArcGIS Pro. I used the point data of where their patrons are located and a 300x300 ft grid to do a spatial join. I was then able to use this grid in the Spatial AutoCorrelation tool to determine if the patrons are clustering and at which distance they experience the highest clustering. I determined that the patrons are clustering most at a distance of 3400 feet.

Results

Application and Reflection

In solving this problems, I learned skills on spatial pattern analysis that I wil be able to take into my future career. Below I describe another scenario where I would need to apply these same skills.

Problem Description

I work for a university. They are interested in knowing if there are any patterns from where their in-state students come from. I would be able to use similar skills to produce a map that would determine and display if there are any clusters.

Data Needed

To solve this task, I would need to be provided with data from the university on where their in-state students are from. I would also need a grid that covers the state.

Analysis Procedure

I would first import the data into ArcGIS Pro. I would then use a spatial join to join the data provided by the university with the grid that covers the state. From there, I would perform a spatial autocorrelation analysis to determine if there is clustering and at what distance it is highest. Finally, I would be able to use the results to to create a map layout for the university.

Return to Top