Alcoholism is a pernicious addiction, with a high rate of recidivism. If we can understand the associations between access to alcohol and violence, we will be better equipped to develop effective strategies and interventions to lessen the costs alcoholism exacts on those who consume it, their victims, and our communities. To get an estimation for lengths of buffer radii, we used the NEAR function, found in Arc Toolbox. This tool calculated the distance from each crime to the nearest alcohol outlet. Buffer Construction We set our buffer thresholds according to the distribution of distances from each crime to the nearest alcohol outlet (see figure above): the median value (966 ft.), the mean (1,403 ft.), and the mean plus one standard deviation (2,844 ft.). Problem Statement: To investigate spatial relationships between access to alcohol and violence in Sacramento, California. Access To Alcohol and Violent Crimes—A Spatial Analysis Sacramento, California Crimes Per Capita by Census Block Group, With Alcohol Outlet Locations Small Buffer Radius = 966 ft. Medium Buffer Radius = 1,403 ft. Large Buffer Radius = 2,844 ft. Study Area 3,945 Violent Crimes Within Study Area Within 1,403 ft. Buffered Areas: 2,515 Violent Crimes 63.8 % of Total Within Study Area 0.025 Violent Crimes/Capita Close-up: Buffer – 1,403 ft. Within 966 ft. Buffered Areas: 1,970 Violent Crimes 49.9 % of Total Within Study Area 0.027 Violent Crimes/Capita Close-up: Buffer – 966 ft. In some cases, as shown in the close-ups above, crime density increased with smaller buffers around the alcohol outlets. However, this was not the case throughout much of the study area. Violent Crimes, Within Total Percent (%) 966 ft. buffers 1,970 49.9 1,403 ft. buffers 2,515 63.8 2,844 ft. buffers 3,445 87.3 Entire Study Area 3,945 100 We looked at whether there was a significant relationship between the occurrence of violent crimes and distance from alcohol outlets. When number of crimes was considered as a function of distance from alcohol outlets, statistical analysis yielded R = 0.266, R Squared = 0.071). We normalized crime as a function of population (we divided the number of crimes in each polygon by the population in that same polygon). The results yielded crime per capita vs. distance from alcohol outlets (see chart at left), with R = 0.042 and R Squared = 0.002. Acknowledgements: Work done by (in alphabetical order): Michael Cozzi, Margarita Kloss, Scott McCarey, and Alexander Ngo Thanks to: Prof. John Radke for his help and guidance, Bridget Freisthler—our client, and the Sacramento Police Department for providing the crime data. More thanks goes out to our classmates, & Juergen, Patty, and Mu Lan. Results, and Some Further Thoughts… Linear regression yielded low associations between violent crime and distance to alcohol outlets. However, the methods employed in our study are by no means an exhaustive approach to the understanding the spatial relationships between violent crime and access to alcohol. Similar analytical methods employing Geographic Information Systems can be used to identify problem areas where alcohol availability does have a strong relationship with assaults. We suggest the following: Buffer around alcohol outlets based on street network distances and travel times, rather than radial distances. In reality, people generally use transportation network to access alcohol outlets. Each segment of the network has different friction factors associated with it and would affect alcohol consumption differently. (For example, highways have a much lower friction factor than residential streets.) Consider variation in police crime data—i.e. between-precinct differences, crime reporting methods, etc. Conduct study by neighborhoods, and see if the neighborhoods themselves account for much of the variation. Use variable buffer distances scaled to population densities rather than distance. Cleaning & Processing Buffer Analysis Results Data Acquisiti on Approach Our project was to explore the spatial relationships between violent crimes and access to alcohol. We analyzed frequency and density of violent crimes as a function of distance from alcohol outlets. To do this, we created buffers of various lengths around the alcohol outlets. We then evaluated whether there were significant differences between the frequencies and rates of violent crime within the buffers of various lengths vs. throughout the entire study area. Identity/Frequency Problem: Crime and alcohol outlets were point data, and we needed counts and densities within each polygon. In addition, some addresses had multiple crimes-- sometimes in excess of 35; we needed a solution that would count multiple records at a single location. Solution: We used the Identity function in ArcToolbox, and intersected the crime data with the census block group polygons. The Identity function assigned the census polygon IDs to each crime point that was within its area. We then used the Frequency function in ArcWorkstation to count the "frequency" of occurrences for crimes in each polygon. Crime densities (as "crime per capita“) could then be calculated by dividing by the number of residents in each polygon. Cleaning and Processing Geocoding Problem: Many of the crimes and alcohol licenses landed on boundary streets for census block groups (even when they had left/right street information in their attribute tables. Solution: Geo-coding was redone, with crimes and alcohol outlets offset 50 feet from the boundary streets' centerlines. This located them within the census blocks, so that they could be counted, buffered, etc. Buffer Analyses We created buffers of varying distances (944, 1,402, and 2,844 ft.) around each alcohol outlet. We intersected the buffer coverages with the census data. For each polygon we recalculated the area, number of crimes, and population. We then compared crime frequencies and rates within the buffered areas, as a function of distance. We also compared crime in buffered areas to the frequencies and rates over the entire study area. The figure below shows the study area, superimposed with all three buffers. Alcohol Outlets Source: State of California, Alcohol Beverage Control Date: 2000 Description: Liquor licenses with street addresses, broken down into off-premises sales (i.e. liquor & convenience stores, and other outlets where alcohol can be purchased), and on-premises consumption, further categorized as bars, restaurants, and pubs. We did not include the restaurant data. Demographics Source: TIGER Census Data Date: 1990 Description: Demographic information, such as: race, ancestry, migration, language spoken at home/ability to speak English, income, poverty status, occupation and employment status, rent/mortgage payment, family composition, etc. Violent Crime Source: Sacramento Police Department, Crime Statistics Date: 2001 Description: Records of crimes, including robbery, rape, narcotics, public drunkenness, DUI; addresses provided for each incident. We selected a subset of the crime database, including assault, battery, homicides, assault with a deadly weapon. Data Sources Number 3,945 Mean 1,403.1 Standard Error of Mean 22.9 Median 966.6 Standard Deviation 1,441.4 Mean + 1 StdDev 2,844 Range 9,206.2 Minimum 0 Maximum 9,206.2 At right is the cumulative histogram for the distance from violent crimes to nearest alcohol outlet. The range of buffer distances we used encompasses the distances with the greatest rate of change in crime frequency, and 87% of total crimes within the study area. 87.3% Variable Variable R R Squared Crime Frequency Distance From Alcohol Outlet 0.266 0.071 Crime Per Capita Distance From Alcohol Outlet 0.042 0.002 Crime Per Capita Alcohol Outlets Per Capita 0.372 0.138 Alcohol Outlet Legend Crimes Per Capita Close-up: Buffer – 2,844 ft. Within 2,844 ft. Buffered Areas: 3,445 Violent Crimes 87% of Total Within Study Area 0.018 Violent Crimes/Capita City & Regional Planning 255, Spring 2002. Department of Landscape Architecture & Environmental Planning University of California, Berkeley http://www-laep.ced.berkeley.edu/classes/cp255/sp2002/cp255-aa/project/present/ Distance from Alcohol Outlet (ft.) Crimes per Capita (# of crimes) Distance from Alcohol Outlet (ft.) Number of crimes
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Alcoholism is a pernicious addiction, with a high rate of recidivism. If we can understand the associations between access to alcohol and violence, we will be better equipped to develop effective strategies and interventions to lessen the costs alcoholism exacts on those who consume it, their victims, and our communities.
To get an estimation for lengths of buffer radii, we used the NEAR function, found in Arc Toolbox. This tool calculated the distance from each crime to the nearest alcohol outlet.
Buffer Construction
We set our buffer thresholds according to the distribution of distances from each crime to the nearest alcohol outlet (see figure above): the median value (966 ft.), the mean (1,403 ft.), and the mean plus one standard deviation (2,844 ft.).
Problem Statement:To investigate spatial relationships between access to alcohol and violence in Sacramento, California.
Access To Alcohol and Violent Crimes—A Spatial AnalysisSacramento, California
Crimes Per Capita byCensus Block Group, With Alcohol Outlet Locations
Small BufferRadius = 966 ft.
Medium BufferRadius = 1,403 ft.
Large BufferRadius = 2,844 ft.
Study Area
3,945 Violent CrimesWithin Study Area
Within 1,403 ft. Buffered Areas: 2,515 Violent Crimes 63.8 % of Total Within Study Area 0.025 Violent Crimes/Capita
Close-up:Buffer – 1,403 ft.
Within 966 ft. Buffered Areas: 1,970 Violent Crimes 49.9 % of Total Within Study Area 0.027 Violent Crimes/Capita
Close-up:Buffer – 966 ft.
In some cases, as shown in the close-ups above, crime density increased with smaller buffers around the alcohol outlets. However, this was not the case throughout much of the study area.
Violent Crimes, Within Total Percent (%)
966 ft. buffers 1,970 49.91,403 ft. buffers 2,515 63.82,844 ft. buffers 3,445 87.3
Entire Study Area 3,945 100
We looked at whether there was a significant relationship between the occurrence of violent crimes and distance from alcohol outlets. When number of crimes was considered as a function of distance from alcohol outlets, statistical analysis yielded R = 0.266, R Squared = 0.071).
We normalized crime as a function of population (we divided the number of crimes in each polygon by the population in that same polygon). The results yielded crime per capita vs. distance from alcohol outlets (see chart at left), with R = 0.042 and R Squared = 0.002.
Acknowledgements:Work done by (in alphabetical order): Michael Cozzi, Margarita Kloss, Scott McCarey, and Alexander Ngo
Thanks to: Prof. John Radke for his help and guidance, Bridget Freisthler—our client, and the Sacramento Police Department for providing the crime data. More thanks goes out to our classmates, & Juergen, Patty, and Mu Lan.
Results, and Some Further Thoughts…Linear regression yielded low associations between violent crime and distance to alcohol outlets. However, the methods employed in our study are by no means an exhaustive approach to the understanding the spatial relationships between violent crime and access to alcohol. Similar analytical methods employing Geographic Information Systems can be used to identify problem areas where alcohol availability does have a strong relationship with assaults. We suggest the following:
Buffer around alcohol outlets based on street network distances and travel times, rather than radial distances. In reality, people generally use transportation network to access alcohol outlets. Each segment of the network has different friction factors associated with it and would affect alcohol consumption differently. (For example, highways have a much lower friction factor than residential streets.)
Consider variation in police crime data—i.e. between-precinct differences, crime reporting methods, etc. Conduct study by neighborhoods, and see if the neighborhoods themselves account for much of the variation.
Use variable buffer distances scaled to population densities rather than distance.
Cleaning & Processing
BufferAnalysis
Results
Data Acquisition
ApproachOur project was to explore the spatial relationships between violent crimes and access to alcohol. We analyzed frequency and density of violent crimes as a function of distance from alcohol outlets. To do this, we created buffers of various lengths around the alcohol outlets. We then evaluated whether there were significant differences between the frequencies and rates of violent crime within the buffers of various lengths vs. throughout the entire study area.
Identity/FrequencyProblem: Crime and alcohol outlets were point data, and we needed counts and densities within each polygon. In addition, some addresses had multiple crimes--sometimes in excess of 35; we needed a solution that would count multiple records at a single location.
Solution: We used the Identity function in ArcToolbox, and intersected the crime data with the census block group polygons. The Identity function assigned the census polygon IDs to each crime point that was within its area. We then used the Frequency function in ArcWorkstation to count the "frequency" of occurrences for crimes in each polygon. Crime densities (as "crime per capita“) could then be calculated by dividing by the number of residents in each polygon.
Cleaning and ProcessingGeocodingProblem: Many of the crimes and alcohol licenses landed on boundary streets for census block groups (even when they had left/right street information in their attribute tables.
Solution: Geo-coding was redone, with crimes and alcohol outlets offset 50 feet from the boundary streets' centerlines. This located them within the census blocks, so that they could be counted, buffered, etc.
Buffer AnalysesWe created buffers of varying distances (944, 1,402, and 2,844 ft.) around each alcohol outlet. We intersected the buffer coverages with the census data. For each polygon we recalculated the area, number of crimes, and population. We then compared crime frequencies and rates within the buffered areas, as a function of distance. We also compared crime in buffered areas to the frequencies and rates over the entire study area.
The figure below shows the study area, superimposed with all three buffers.
Alcohol OutletsSource: State of California, Alcohol Beverage ControlDate: 2000Description: Liquor licenses with street addresses, broken down into off-premises sales (i.e. liquor & convenience stores, and other outlets where alcohol can be purchased), and on-premises consumption, further categorized as bars, restaurants, and pubs. We did not include the restaurant data.
Demographics Source: TIGER Census DataDate: 1990Description: Demographic information, such as: race, ancestry, migration, language spoken at home/ability to speak English, income, poverty status, occupation and employment status, rent/mortgage payment, family composition, etc.
Violent CrimeSource: Sacramento Police Department, Crime Statistics Date: 2001Description: Records of crimes, including robbery, rape, narcotics, public drunkenness, DUI; addresses provided for each incident. We selected a subset of the crime database, including assault, battery, homicides, assault with a deadly weapon.
Data SourcesNumber 3,945Mean 1,403.1Standard Error of Mean 22.9Median 966.6Standard Deviation 1,441.4Mean + 1 StdDev 2,844Range 9,206.2Minimum 0Maximum 9,206.2
At right is the cumulative histogram for the distance from violent crimes to nearest alcohol outlet. The range of buffer distances we used encompasses the distances with the greatest rate of change in crime frequency, and 87% of total crimes within the study area.
87.3%
Variable Variable R R SquaredCrime Frequency Distance From Alcohol Outlet 0.266 0.071Crime Per Capita Distance From Alcohol Outlet 0.042 0.002Crime Per Capita Alcohol Outlets Per Capita 0.372 0.138
Alcohol OutletLegend
Crimes Per Capita
Close-up:Buffer – 2,844 ft.
Within 2,844 ft. Buffered Areas: 3,445 Violent Crimes 87% of Total Within Study Area 0.018 Violent Crimes/Capita
City & Regional Planning 255, Spring 2002.Department of Landscape Architecture & Environmental PlanningUniversity of California, Berkeley http://www-laep.ced.berkeley.edu/classes/cp255/sp2002/cp255-aa/project/present/
Distance from Alcohol Outlet (ft.)
Crim
es p
er C
apita
(# o
f crim
es)
Distance from Alcohol Outlet (ft.)
Num
ber o
f crim
es
Related Documents
