Format report of viewing crime in Cleveland neighborhoods through Spatial Statistics in the ArcGIS program.
Breaking Down Crime in Cleveland Neighborhoods Using Spatial Statistics
Hasani Wheat Cleveland State University Levin College of Urban Affairs PDD 643- Advanced GIS Dr. Sung-Gheel Jang December 16, 2011
Contents Abstract..2 Background3 Goals and Objectives.............................................5 Data6 Methods and Analysis8 Discussions and Summary...12 References19 Appendix..20 1. 2. 3. 4. Attribute Tables...20 Maps.........21 Results..27 Data Dictionary....30
Abstract The association of Cleveland, Ohio with the presence of crime has been an ongoing topic of conversation for both residents and non-residents for many years. In Cleveland, some neighborhoods have a bad reputation of being dangerous areas because of foreclosures, high vacancy rates, and other housing and demographic information. Oftentimes, some neighborhoods that are viewed as unsafe because of their perceived high amounts of crime are not as unsafe as perceived when data is presented. Another problem that is noticed about crime in Cleveland neighborhoods is its lack of availability and detail. While there are maps that analyze crime in the Cleveland area, most maps do not show the most detailed scale of analysis, which is through Census block groups. Additionally, there are few maps available that display crime as current as 2010. This paper will detect hot and cold spots in Cleveland neighborhoods using crime data from the Cleveland Police Department via NEO CANDO. Taking it a step further, this paper will also look at the outliers that are near or within those hot and cold spot clusters. Identifying where these outliers are located and in which Cleveland neighborhood will be critical for conducting further research and analysis for showing crime. Using the Census block groups instead of a less detailed boundary files such as Census tracts will allow for more data to be clustered for the hot spot analysis. The Census block group data can also be used to establish where the outliers of crime types are in Cleveland neighborhoods.
Background This project that focuses on using spatial statistics to display crime in Cleveland neighborhoods is important because the maps and data will make it easier to see the concentration of crime in a detailed manner. The perception that Cleveland is an unsafe place due to its crime is prevalent in the minds of many people including individuals that are from Cleveland neighborhoods. In order to change the minds of Cleveland residents, a visual representation that reflects the types of crimes that are present in Cleveland neighborhoods needs to be presented to the urban masses. The problem statement for this project is: Is there a statistically significant relationship between the number of crime counts of a particular type and the location of a particular census block group that lies within a particular neighborhood? As mentioned in the proposal of this project, in order to get a better understanding of the levels of crime in Cleveland, there must be a categorical breakdown of where crime is located in Cleveland and more specifically, its neighborhoods. Although the objective is to break down any myths associated with crime types in a given Cleveland neighborhood, graphical representation is also critical. Cleveland as a whole is not a dangerous place but by providing maps of what particular crimes affect what particular neighborhood or even what particular crimes do not affect a particular neighborhood, the hope is that the maps will alleviate fears that people have about that particular crime. For this project, there are only a couple of maps that are presented; however, these maps will be a small representation of the type of crime and where crime is located. In a nutshell, the problems that the project will attempt to address are 1.) Where is a particular crime type located in Cleveland with hot and cold clustering? 2.) Where are the outliers in a hot or cold cluster? 3.) How to assess the results of the hot and cold clustering as well as the outliers within these clusters in a real life situation? The audience should expect three things from this project: a visual representation of whether a particular type of crime is either a high crime rate cluster, a low crime rate cluster, or an area that falls in between the extreme values of crime located in each Cleveland neighborhood represented by Census block groups, an explanation to the hot and cold spots results of a particular crime for each neighborhood, and an explanation of where the outliers are within or near a cluster of high or low crime. Once these methods that analyze various crimes are created, an explanation of criminal activities in a spatial context can be established by comparing socioeconomic factors such as vacancy and foreclosure rates to each other (Greenburg & Rohe, 1984).
Easy access to retrieve the information so that people can see the data would be helpful so that people can be aware that the data exists. Policymakers and city officials will be able to seek out the approximate areas of where the crime is located. For example, according to the Cleveland Police Department, the North Broadway neighborhood has one of the highest average rates of violent crime out of the 36 Cleveland Statistical Planning Areas. By selecting the area that has the highest rate of crime, the policymakers will be able to extract this location from the rest of the dataset, zoom in on the location, compare it with the Statistical Planning Area Map, study the area, and make recommendations on as to how to reduce the crime count (See Table 1, Maps #1 & #2). Some of the research that was influential to the development of this project was Extend Crime Analysis with ArcGIS Spatial Statistics Tools by Lauren Scott and Nathan Warmerdam, Mapping Crime: Understanding Hot Spots by the U.S. Department of Justice- Office of Justice Programs, and the Rebuilding Blocks article by Randall McShepard & Fran Stewart from PolicyBridge. The Scott and Warmerdam article provided me with an introduction to the importance of creating hot and cold spots using statistical analysis to easily depict the data. The Mapping Crime article indicates the importance of establishing theories to help support the data shown in the maps as to why the high or low crime areas are the way they are and how policymakers and other people of interest will be able to make decisions based off these theories. The Rebuilding Blocks article addresses the fear and the perception of crime in Cleveland and its neighborhoods that is mentioned earlier in this paper. As for the cluster and outlier analysis portion of the project, there are also some research and literary works in which I referred to in assisting me with this project. A 2007 course project titled Drug arrests in High Gun Crime Locations of Dallas Texas by Josh Taylir focuses on how gun crimes are either spatially clustered, random, or dispersed using the Local Morans I statistical tool. The other article that helped me to understand the importance of cluster and outlier analysis is, ironically, named after the statistical tools namesake, Luc Anselin. The article, Review of Cluster Analysis Software, addresses the importance of the Anselin Local Morans I statistical tool. Some of the programs mentioned in the review are CrimeStat and GeoDa, which are two recognizable programs which evaluate maps on a statistical basis. The importance of this article is that many statistical programs utilize the Local Morans I as a tool; ArcGIS takes it a step further by using Local Morans I and incorporating the statistical tool into
what will be a valuable asset in viewing a map showing clusters and outliers of crime in a Cleveland neighborhood. Additionally, in the PowerPoint created by Scott and Warmerdam titled Spatial Statistics for Public Health and Safety, Scott and Warmerdam further define cluster and outlier analysis as gaining a better understanding of feature distribution through degree of clustering or dispersion across study area (Scott and Warmerdam, Spatial Statistics).
Goals and Objectives The goal of this project is to analyze the locations of crime types in the Cleveland neighborhoods through Spatial Statistic tools. These Spatial Statistic tools are the Hot Spot Analysis, which uses the Getis-Ord Gi* statistic and the Cluster and Outlier Analysis, which uses the Anselin Local Morans I statistic. By creating a visual of where crime types are located in the Cleveland area, people will better understand where in Cleveland different crime types are prevalent. The objectives of the project that will help meet the goal are to identify which Cleveland neighborhoods have high counts of crime (hot spots that are statistically significant), Cleveland neighborhoods that have low counts of crime (cold spots that are also statistically significant), as well as Cleveland neighborhoods that have crime counts that average in the middle (the majority of census block groups that are within range of the mean). Additionally, the other objective to this project is to seek out Census block groups in a high or low cluster that have a higher or lower crime number than the rest of the cluster. In other words, the analysis will be similar to that of the hot spots; however, the census block group within or near a cluster of low crime will have a higher than usual number than its surrounding census block group. This also happens to be true with high crime cluster that have a couple of census block groups that have lower than expected crime numbers.
Data Looking at the breakdown of specific crimes in Cleveland from the 2010 crime reports from City Data (http://www.city-data.com/crime/crime-Cleveland-Ohio.html), I wanted to use a variable that was deemed a common problem in Cleveland neighborhoods and a serious but infrequent crime problem. Th