Cluster Analysis

Analysis of Understanding Local Spatial Patterns Focused on Two Different Real-World Scenarios

Problem and Objective

Spatial patterns analysis examines the clustering, random, or dispersal pattern of the distribution of attributes. It provides insights into the monitoring of conditions, tracks changes through time, and compares features over space and time. The purpose of this analysis is to identify clusters in false alarm incidents based on feature locations and ranking of calls in the battalion 2 boundary and examine the hotspots and cold spots based on feature locations and level of median household income in Dallas county for job creation.

Analysis Procedures

Strategies: The strategies adopted for this analysis are explained as follows. I used ArcGIS Pro (Version 2.9) software. I used the following Modules/tools for the analysis: Cluster and Outlier Analysis (Anselin Local Moran’s I) and Hot Spot Analysis (Getis-Ord Gi*) tool. Data and data types – The data include (a) an mxd file of tutorial-exercise layout containing Battalion 2 layer, incident – 2015 feature class, Census 2010 population layer, and active and proposed stations feature class (b) an mxd file of tutorial-exercise layout map containing the major road feature class and Dallas County census block feature class. Data sources – GIS 520 Fall Semester, NCSU, and DavidW. Allen (2016).

Methods: I imported the first exercise map layout (mxd file) of Batallion 2. I opened the attribute table of the incident of the January 2015 layer. After searching for the Cluster and Outlier Analysis (Anselin Local Moran’s I) tool, I used this tool for the cluster and outlier analysis of the incident of false alarm in January 2015 using an inverse distance spatial relationship. I reran the tool once again using a fixed distance band (900 feet) with aim of controlling the search distance it uses to find the significant number of neighbors. The 900 feet distance is the distance at which the z-score peaked in the previous analysis of global spatial pattern analysis. I changed the symbology accordingly. Next, I imported the second exercise map layout for the Dallas County Census blocks. I obtained the high-income and low-income clusters by running the Hot Spot Analysis tool using the Median Household income field and distance band of 5280 feet. I changed the symbology accordingly.

Workflow diagram of the analysis

Results

The map above shows the result of the cluster and outlier analysis of the call for service in Dallas County at an inverse distance band. The areas of high values cluster are symbolized in red, areas of low values cluster are symbolized in blue, and outliers are symbolized in yellow. The result was displayed over the median household income to visually examine if any underlying factor exists.

The map above shows the result of the cluster and outlier analysis of the call for service in Dallas County at a fixed distance band. The areas of high values cluster are symbolized in red, areas of low values cluster are symbolized in blue, and outliers are symbolized in yellow. The result was displayed over the median household income to visually determine whether there is any relationship between both.

The map above shows the result of the hotspot analysis of the pattern of median household income distribution in Dallas County at a distance of 5280 feet (1 mile). The areas symbolized in blue represent areas of clustering of low values and those in dark red are areas of clustering of high values and they are significant at a 99% confidence level.

Application & Reflection

Problem Description: The coronavirus outbreak spread all over the 50 states with over a million confirmed cases across the nation. Tracking the trend in this spread will help access the rate of the spread of the virus, especially those people already tested positive. The purpose of this analysis is to examine the places with a high-value and a low-value number of cases of COVID-19 in North Carolina in 2020.

Data Needed: The data needed include (a) the study area (North Carolina) boundary, (b) the point feature of individual cases of people who tested positive, and (c) population data of NC state.

Analysis Procedures: I will add all feature class data to the map. I will use the Cluster and Outlier Analysis (Anselin Local Moran’s I)tool to perform the cluster and outlier analysis of the incident of cases of COVID-19 in 2020 using an inverse distance spatial relationship. I will also run the tool again using a fixed distance band. This fixed distance band to be used would have been obtained from the previous analysis of high/low clustering analysis using the high/low clustering analysis tool. I will then use the graduated color symbol for the population layer. I will visually examine the relationship between the clustering of COVID-19 cases and the population distribution in North Carolina. I will obtain the high number (hot spots) of confirmed cases and the low number (cold spots) of confirmed cases clusters by running the Hot Spot Analysis tool using the frequency of confirmed cases in each county in North Carolina. I will change the symbology and interpret the results accordingly.

Page updated