Govt Change | №2 Mapping Police & Crime in Philadelphia
Examining the Relation between Police Station and Sexual/Child Crime with Other Factors Following Last Article
Introduction
Following my MSSP 631 Law and Social Policy final paper about Child Sexual Abuse (CSA) Justice, I’m curious about two questions: 1) how relevant crime has been distributed around police stations and 2) how socioeconomic factors affect such distribution. To choose a location for mapping, Philadelphia is the most familiar for me, where has published crime data. Despite limited government data collection and publishing compared to other states, Pennsylvania has constantly promoted child welfare legislative reforms since amending the definitions of child abuse and perpetrator in 2014. The most recent legislative reform is that Pennsylvania has become the third state to ban child marriage by adding 18 as the minimum marriage age to the bill.
According to the Pennsylvania Department of Human Services, the Philadelphia county has 945 substantiated child abuse reports in 2018, which 483 CSA reports are the second- highest allegations besides physical abuse. The most prevalent relevant crimes in Philadelphia are rape, sexual assault, aggravated sexual assault, and sexual acts or attempt to commit a sexual act upon a minor. Since an individual can be charged with many different sex crimes, the Philadelphia Police Department attributes CSA to both child abuse and sexual abuse. In Pennsylvania Criminal Law, both sexual offenses and offenses against the family define terms related to CSA.
Mapping Evolution
Crime mapping has long been the fundamental past of crime analysis, which can be shadowed as “pin on maps” back to at least the years 1900. Modern forensic geography was formalized in the 1970s as applied geography. At that time, geographers acted as witnesses and consultants and investigate crime, legal disputes, or cases. Geographical Information Systems (GIS) is one kind of geoforensics, which also includes pedology, mineralogy, petrology, geophysics, natural geography, remote sensing, and location data.
In 1999, one of the first GIS practice manuals of mapping crime was published. At that time, GIS technology was mainly to “help analyze information”, such as types of crime by location, victim population groups served and underserved, and the location of victim service organizations and their geographic service areas. This information can be used to 1) examine the availability of basic services and the sufficiency of services for specialized population groups; 2) display multiple funding sources in a geographic area to help in the fair distribution of resources; 3) develop a strategic program and financial plans for the maintenance and development of victim services. In recent years, crime analysis has turned into a broad-spectrum term, in which new technologies include identifying hot spots, statistical regression and correlation, and predictive analytics. Those policing strategies reduced crime rates by allowing departments to concentrate officers on troublesome areas instead of randomly patrolling.
Data & Methodology
ArcGIS is one of the most common platforms to meet GIS needs initially released in 1999, which can “perform analysis on spatial data, manage a large amount of spatial data and produce cartographically appealing maps that help with decision making”. This ArcGIS model contains three datasets.
- Crime Incidents
Crime incident dataset comes from the City of Philadelphia this year and is updated every year by the Police Department, which defines 33 types of crimes including violent offenses (aggravated assault, rape, arson, etc.) and non-violent offenses (simple assault, prostitution, gambling, fraud, etc.). Among all 53,022 reports, 416 relate to CSA, which 67.5% is rape and 32.5% is offenses against family and children. Although this dataset doesn’t specify CSA, such a result is similar to 483 CSA reports by the Pennsylvania Department of Human Services in 2018 and helps my estimation. However, this estimation is not mutually exclusive and collectively exhaustive. On the one hand, the number is overestimated since the concatenation of rape and family offenses includes other types of sexual crimes such as adult rape and abuse. On the other hand, the number is underestimated since current reports have updated until May 2020.
2. Police Stations
The police station dataset was uploaded by the City of Philadelphia in 2016 yet doesn’t need to be updated frequently. It includes 24 locations of police stations, which contains not only 21 district police stations distributed within six directions of Philadelphia (east, northwest, northeast, central, southwest, & south) but also 3 non-district police stations (the Police Headquarter, Center City District Police Station, & UPenn Police Station). The procedure to investigate CSA crime includes: 1) receiving the initial report of rape or other sexual assault by the first officer(s) in any police station; 2) taking in charge of the victim by the transporting officer(s); 3) investigating all incidents by the Special Victims Unit (SVU) officer(s) and being overseen by supervisors from different units.
3. Demographics
The demographic dataset comes from Social Explorer’s 2018 American Community Survey, which was collected every year by U.S. Census Bureau. This dataset combines the population, race, and average household income information from 384 census tracts in Philadelphia.
Spatial Analysis
In Figure 1, I displayed the Philadelphia shapefile and changed its coordination projection to the standardized World Geodetic System (WGS) 1984. Then I displayed points of crime data and police stations by their coordination and joined the map with the demographic table by their census tract code. The buffers indicate a 3,643-feet radius around the police stations.
To answer the first half question of Figure 1’s subtitle — whether average household income affects crime — I first used graduated color in symbology to distribute residents with different average household income in Philadelphia and found that the North generally earns lower income than the South and indeed more crimes have happened in red areas with the lowest income. To answer the second half question, I first added a shapefile of ten city council districts to better visualize police station segmentation, which was provided by the Philadelphia City Planning Commission. There are 1–3 stations per district, in which more police stations are located in the South than the North. More police stations seem to lower the crimes, but does that mean average household income and police station distribution affect crime negatively?
Joining the police station table with the crime incident table, I identified how many feet each police station is distant from the closest crime and labeled police stations with its location if the closest crime is within 660 feet, which there are six out of 24 police stations. Among these six stations, only two are in the North and the other four are all in the South, which means crimes are much nearer to the southern police stations despite the lower amount of crimes than the northern stations. The zoom-in circles highlighted two police stations that have closest rape and offense against family and children respectively. Police station with the closest rape is the Police Headquarter at 750 Race St, which is just 170-feet (about 30-second walking) from the crime. And police station with the closest offense against family and children is the 17th District police station at 20th St & Federal St, which is only 15-feet (right near the building) from the crime. Interestingly, both stations are in the South with diverse average household income distribution rather than in the North with the lowest income. Therefore, although we can say that average household income and police station location correlate with the number of crimes in general, both of them don’t correlate with how close the crime can happen.
Besides average household income, I also want to investigate whether other demographic information such as population density and race affects the two crimes closest to the police stations. Therefore, I changed the symbology graduated color to population density and added pie charts to display race distribution in these two locations in Figure 2.
Overall speaking, we can find that more crimes happened in the North with lower density than the South with higher density. Police stations whose closest crimes are within 660 feet are all located in areas with medium-high population density. Although both stations are located in the highest population density area compared to other tracts, the population density of the Police Headquarter is at the edge of this area, which is much lower than that of the 17th District. Race distribution is more diverse at the Police Headquarter, and the 17th District has over three times of African Americans than the Police Headquarter.
The above two maps can depict overall crime and police station distribution and identify stations with the nearest distance to the crimes. But which police stations are closest to severe crimes? Because the buffer distance has a constant width, it can only display discrete data within the specified range but couldn’t help with identifying the severity. Therefore, I created the density map, a popular technique to identify hot spots by the police, to cluster data together with a 10,930-ft radius and a 182-ft cell size in Figure 3.
The 35th District police station at N Broad St and Champlost St is surrounded by the severest crime, following by the 17th District and the 24th District at 3901 Whitaker Ave. Unlike other stations that are usually surrounded by one type of crime, the 35th District’s surrounding crimes include both rape and offenses against family and children. Combing with the last two maps, we can find that six police stations whose closest crime is within 660 feet are all located at a low-medium crime density area and the 35th District is one of them. We can also find that the 35th District is in the low-income North with low population density. All above three maps’ findings are summarized below (Figure 4).
Conclusion
The above summary could answer two questions at the beginning of the paper: 1) 25% of the police stations have crimes that happened two-minute walking away from their stations and 2) socioeconomic factors such as income and population density negatively correlates to the number of crime incidents. A further suggestion for policing strategy is to focus on areas near the 17th District in the South and the 35th District in the North, which could be called as “hot spots” for rape and family/children offenses.
While the paper gives a clear picture of crime and police station distribution, the biggest limitation is the vague definition of CSA in criminal settings and the possibilities of underestimating or overestimating the reports. Similar publishing about Pittsburgh sexual violence, the researcher chose to analyze rape and aggravated assault. Since aggravated assault is defined as an attempt to “cause serious bodily injury to another” in Pennsylvania, it specifies neither child abuse nor sexual abuse so that I eliminate it from the analysis. However, I hope that the Philadelphia Police Department who updates the data or Philadelphia District Attorney’s Office who visualizes the data can specify CSA in the crime dataset or report.
Next-step researches can be categorized as two ways: 1) to benchmark the same techniques on some cities such as Cixi in China where have published CSA-related crime data since my law and social policy paper compare CSA environment in US and China; 2) to deepen CSA research in Philadelphia by adding other factors such as mental health and alcohol/drug usage, mapping Philadelphia child service locations and indicating driving routes from each criminal case to the nearest service center, and even combining with statistical analysis such as correlation and regression.
To prioritize one, I would like to map child service locations in Philadelphia, as “the Commonwealth has a vested interest in the thoroughness and quality of child protection investigations”. Philadelphia has the fourth-largest child welfare and spent 27% of the statewide child abuse expenditures in 2018. Accurate and timely identification of abuse and neglect can prevent victims from experiencing additional maltreatment and perpetrators from accessing additional victims.
To sum up, this paper examined the use of ArcGIS to illustrate the relation between police stations and child/sexual crime with other factors. This model incorporated the City of Philadelphia and US Census data and visually identified hot spots related to child/sexual crime.
References
1 Liu, L. (May 2020). Child Sexual Abuse Justice: An Overview of the US and China legislatin and policy. Medium.
2 29th Year of Child Maltreatment Reporting. (2018). Washington, DC: U.S. Children’s Bureau.
3 Mcnamara, A. (May 2020). Pennsylvania just became the third state to ban child marriage. CBS News.
4 Child Protective Services Annual Report. (2018). Phildelphia, PA: Pennsylvania Department of Human Services.
5 Types of Sex Crimes in Philadelphia. (n.d.). Pennsylvania Criminal Law Practice of Price Benowitz.
6 Rape and Other Sex Offenses, Directive 5.25. (Feb 2013). Philadelphia Police Department.
7 Sexual Offenses, Pa. Cons. Stat. 18 (2019), Chapter 31 (Pa. C. S. A. 2019).
8 Offenses Against the Family, Pa. Cons. Stat. 18 (2019), Chapter 43 (Pa. C. S. A. 2019).
9 Shafique, I., Zahra, S. A., Farid, T., & Sharif, M. (Dec 2017). Role of GIS in Crime Mapping & Analysis. Sukkur IBA Journal of Computing and Mathematical Sciences-SJCMS, 1, 2, 39–47.
10 Harries, K. (1999). Mapping Crime: Principle and Practice. Washington, DC: U.S. Institute of Justice.
11 Stoe, D. A., et al. (Feb 2003). Using Geographic Information Systems to Map Crime Victim Services. Washington, DC: U.S. Institute of Justice.
12 Track Crime Patterns to Aid Law Enforcement. (n.d.). Learn ArcGIS.
13 Id. at 9.
14 Urciuoli, M. (2020). Crime Incidents. OpenDataPhilly.
15 City of Philadelphia Police Department. (2016). Police Stations. Data.World.
16 Districts & Units. (n.d.). Philadelphia Police Department.
17 Id. at 6.
18 Originally set up as 0.01 map unit. 1 WGS84 map unit = 69 miles = 364,320 feet.
19 Patel, D. (2015). City Council Districts. OpenDataPhilly.
20 An average person’s walking speed/distance. (n.d.).
21 Originally set up as 0.03 map unit.
22 Originally set up as 0.0005 map unit.
23 Alder, H. (Dec 2019). Accessibility of Resources for Survivors of Sexual Violence. StoryMaps.
24 Aggravated assault, Pa. Cons. Stat. 18 (2019), Chapter 27 § 2702 (Pa. C. S. A. 2019).
25 Font, S. A., Miyamoto, S., & Pinto, C. N. (Jan 2020). Child Sexual Abuse and Exploitation in Pennsylvania. The Center for Rural Pennsylvania.
26 Moya, E. L. (Dec 2017). Child abuse: This is what Philadelphia is doing to fight it. AL DÍA News.
27 Id. at 4.
