Problem:
i. Fort Worth Fire department in Texas asked for a spatial analysis project and wanted to know if Emergency Service Calls for Battalion 2 have a tendency to cluster By using Average Nearest Neighbor Analysis tool I need to analyze the data and determine the degree of clustering and dispersion.
ii. They need to know the distance at which the clustering occurs. I need to use High Low clustering (General G )tool to look at similarity of the values associated with the features within a critical distance of each other.
iii. Fire department wants me to study the relationship of each feature to the other features of Batelion 2 study area The Multi Distant Spatial Cluster Analysis tool should be used for this study. Using Reply’s K function for pattern analysis measures the distance between features to determine clustering.
iv. Oleander library, TX are asked for patron location analysis by using Spatial Autocorrelation tool to determine Moran’s I. As patron is the point data, I need to aggregate this data with polygon data before running Autocorrelation tool. A 300 ft Grid has been provided for this exercise.
v. I need to show a comparison between the priority incident of Jan 15 z score and cluster types as well as the median household income of Fort Worth, TX.
vi. Dallas county, TX Census 2010 Household income data has been provided. It contains a P053001 value field for median income. I need to create a map that shows the cluster of hot and cold spots of median income values using Hot spot analysis (Getis Ord G). I need to use a fixed distance of 5280 ft.
Analysis Procedures:
i. In order to find the degree of clustering and dispersion in the study area, Nearest Neighbor Analysis tool had to be chosen. There were two different points layers. I needed to turn off the one layer which was not needed to analyze. Only particular Incident data were going to analyse and incident type had to be selected. Euclidian Distance method had to be chosen by default as well as the study area in square feet. By choosing generating report for this process, through the summarized result from Geoprocessing menu, Observed and expected mean distance, nearest neighbor ratio as well as z score had been found out for the data.
ii.In order to solve this problem, I chose the High Low Clustering (General G) analyzing pattern tool from Arc toolbox. Then I chose the appropriate feature for input feature class and appropriate value for input field by using fixed distance band and Euclidean method. The distances to try should be from 1000 ft. to 1400 ft. at 100 ft intervals in General G tool. As map units are feet, these numbers can be entered without adding units. I need to always choose the option to generate report for geoprocessing result.
iii.Locate Multi distance spatial cluster analysis (K function) tool. Input the particular feature class and input the particular weight field, number of distance band of 10, with beginning distance 200 and distance increment of 100, optional confidence envelop of 0, user provided study area feature class for study area method, set boundary correction method is none, study area feature class to the appropriate study area. K function table has been added to the table of contents.
Rerun the Multi distance spatial cluster analysis (K function) tool by using compute confidence envelop of 99 permutations with using all other field same.
iv. Perform a spatial join of the 300 ft grid by using provided point data location. Perform a spatial locational quary. Determine the distance to use the analysis,eg. 300 ft grid represents 2800 ft. Run the spatial autocorrelation tool for the new grid layer using the range of test distances. Record the z score and distances with the confidence levels. And determine if I can reject null hypothesis. Identify and analyze the most significant results.
v. I need to run Cluster and Outlier Analysis tool (Anselin local Moran’s I) by using priority ranking incident value with inverse distance and Eucladian method and 0 prominent. So the z score are found.Then I need to make a cluster type layer to show high low cluster type of incidents. Census block groups are classified as graduated colors for median household income. Then I found out high and low clustered type of incidents are located in lower median household income area. Again I need to run Cluster and Outlier Analysis tool(Anselin local Moran’s I) with fixed distance band of 900 using same incident value.
vi. In order to solve the problem, main data has to be used as input feature, target value and fixed distance has to be chosen to run the Hot spot analysis tool(G.O.G).
Dallas county income data is used as input feature, median value of P052001 as input field, fixed distance of 5280 ft has been chosen to run the Hot spot analysis tool(G.O.G). After running the tool, hot spots/red color are visualized in the higher income area and inversely cold spots /blue color are visualized in the lower income belt.
Workflow Diagram
Result:
Application & Reflection:
i.Average nearest neighbor analysis (clustering by location) looks at each feature and single nearest feature and then calculate mathematical index.It is used to see if the physical locations are closer together than would be expected with a random distribution.
I can use this tool to find out the Spatial pattern of the houses and crime reports of an area, Then I can compare the spatial pattern of houses and crime report for that area.
ii.Getis Ord General G (clustering by value) statistic looks at the similarity of the values associated with the features within a critical distance of each other. It is used to see if there are other areas in which similar values are closer together than would be expected with a random distribution.
I can compare the spatial pattern of different type of commercial center like Car dealers and fitness center within a city, look for which types cluster has highest competition spike.
iii. Multidistance Clustering (Reply’s K function)(clustering by location)is used to determine if physical location are clustering due to factors beyond the next nearest feature. Ripleys K function can be used to study the cluster of certain species in certain area.
iv.Spatial Autocorrelation(Global Moran’s I)(clustering by both location and value) is used when the physical location data has an attribute associated with it that may be influencing the clustering. This tool may be used for gas production level.
v.Clustering and Outlier analysis(Anselin Local Moran’s I) is used when someone wants a graphic output on the map and clustering is due to both location and associated feature with it. Crime rate analysis can be established by using crime data.
vi. Gi Hot spot Analysis tool can be used for any disease outbreak investigation purpose. We need to know the area first, how many cases has been identified. Then I need to create weighted data points for individual case. I need to use collect events tools from Utilities tool set. Then Hot spot analysis can be performed by using count as input value. The larger the z score is, the intensity of diseases are more profound.