Information System
1
Visualizing Crime in Philadelphia
Figure 1 – The city of Philadelphia divided into police districts
Abstract - Philadelphia has consistently ranked as one of the most dangerous cities in the United
States. The City of Brotherly Love is plagued by violent crime and its distinction as the most
impoverished among America’s 10 largest cities. There are a number of factors that contribute to
these issues. This research aims to visualize the breadth and depth of crime in Philadelphia in
order to gain a better understanding of what specific types of crimes occur in the city and where.
1. Introduction
The crime rate across the United States as a
whole has fallen significantly over the past
20 years. The Federal Bureau of
Investigations (FBI) releases an annual
report titled Crime in the United States
(CIUS). This report shows that
Philadelphia’s crime rate has fortunately
followed that same pattern as well.
According to the FBI’s Uniform Crime
Reporting (UCR), the number of violent
crimes reported to the Philadelphia Police
Department dropped from a peak of 23,031
in 1999 to 15,925 in 2014 (the most recent
year available) [1]. The department also
highlights crime statistics in their own UCR.
2016 saw the fewest number of violent
crimes since 1979, fewest property crimes
since 1971 and the fewest number of
robberies since 1969 [2]. Nonetheless, the
numbers do not tell the full story of what is
happening in the city.
The specific types of crimes that are
committed, as well as where in the city they
occur exactly, provide more meaningful
grounds for visualizing the numbers behind
the data. The crime rate is a common
misperception amongst citizens of the
2
United States. Even though data shows that
crime has fallen significantly in recent
times, the majority of people, including
prominent politicians, actually believe that
crime has gotten worse [3]. The gap that
exists between the reality of people’s safety
with their oft misconstrued perception of
their safety was a major motivation behind
our research. We aim to visualize the data in
a way that provides a more clear and
complete picture of crime in Philadelphia
with a focus on gun violence.
We obtained all our data sets from Open
Data Philly, the City’s official open data
repository. The information included in
these data sets that we used in our analysis
includes crime incidents, shootings and
police districts. The data fields included
information such as the date of the crime,
location of the crime, specific type of crime
(with a description and code), the police
district the crime was committed in, among
other details of the crime.
Our expectation, like the majority of the
public, was that crime in Philadelphia has
been increasing in the past decade. We
anticipated our analysis to show that
increase. Additionally, we suspected that the
poorest city districts would have higher
crime rates than other parts of the city.
Lastly, we expected to find a large number
of gun violence across the city.
The paper is organized into five additional
sections. The next four sections will each
include one visualization that we produced.
Each visualization is broken down into the
data used to generate it, the results that are
shown and our interpretation of the results
obtained. We will conclude our findings in
the final section of our paper.
2. Visualization I – Crime in 2016
Figure 2 – Number of Incidents by Type in 2016
3
2.1 Data
The data set used to produce this
visualization spans from the years 2006 to
2016. It includes every crime that was
reported to the Philadelphia Police
Department during that time. Other elements
of the crime data include the date and
location of the crime, the police district it
occurred in and a brief description of the
crime. This visualization includes data from
only 2016. The main data element that we
analyzed in this visualization was the
specific types of crime that were committed
during that year. The goal was to visualize
what are currently the most common threats
to public safety for Philadelphia residents.
2.2 Results
The first noteworthy result from this
visualization is that by far the most common
type of crime reported is categorized under
“all other offenses”. This means that these
crimes could not be categorized into any of
the other types of crimes. The top five
specific types of crimes are: narcotic/drug
law offenses, assaults, thefts,
vandalism/criminal mischief and recovered
stolen motor vehicle. The number of
occurrences of these top five crimes alone
total over 10,000 incidents.
2.3 Interpretations
The first crime type categorized as “all other
offenses” tells us a couple of different
things. The first is that these crimes were
random since they could not be categorized
as any of the other major types of crime
included in the list. Secondly, it tells us that
there were a large number of these kinds of
crimes, given how many there were. The
ambiguity that this categorization contains
makes it difficult to come up with any other
determinations related to its content.
The top five specific types of crimes listed
provides a much richer picture because of
how specific they each are. By analyzing
these categories, it is clear that Philadelphia
has a rampant drug problem. There were
over 3,000 drug offenses reported in 2016,
an average of about 8.5 incidents per day.
This is the reality that many Philadelphia
neighborhoods, such as those in North
Philadelphia, Fairhill and Kensington, face
on an everyday basis. Assault, theft and
vandalism/criminal mischief are also daily
occurrences in the city.
Interestingly, the City does not include
shootings in this data set as a specific type
of incident reported, although it does include
other crimes that involve guns such as
aggravated assault, robbery and homicide by
a criminal with a firearm. The reason is that
Open Data Philly actually has a separate
data set exclusively for shooting incidents,
which we explored and analyzed in
Visualization 3.
3. Visualization II – Historical View
of Crime in Top Five Districts
4
Figure 3 – Top Five Districts All Crime, 2006-2017
3.1 Data
We produced this visualization using the
same data set as the first. The only
difference is that the main data element we
analyzed here was the total number of
crimes in each district by year, spanning
from 2006 through 2017. 2017 is not over
yet, so it is the only year that is not entirely
complete. There are 21 districts in
Philadelphia, but our goal here was to
visualize the five most dangerous ones and
put them into historical context over the past
decade.
3.2 Results
Districts 12, 15, 19, 24, and 25 rank as the
most dangerous areas of the city over this
time period. It is important to mention that
the data set only includes crimes that were
reported. Many crimes are not reported to
police for a variety of reasons such as the
lack of trust and cooperation with police
within the community. This should be taken
into consideration since the most crime-
ridden areas are often prone to
underreporting. This visualization shows
that the total number of crimes reported in
the top five districts has steadily decreased
5
over the past decade. It dropped from a peak
of about 77,000 in 2006 to just under 60,000
in 2016. This is a 22% decline in crime over
the past decade.
This is the bigger picture, but by looking at
the numbers, it is also apparent that the total
number of crimes for all five districts have
not decreased every year. Years 2012 and
2014 actually saw an uptick in crime.
Additionally, none of the districts show a
steady yearly decrease. They all fluctuate.
The largest fluctuation appears to be with
District 25 where under 11,000 incidents
were reported in 2016 compared to over
18,000 in 2006.
3.3 Interpretations
Our findings contradict our initial
expectation that crime has gotten worse in
Philadelphia over the past 10 years. Even
though this visualization shows the steady
decrease in crime for only the top five
districts, further analysis of the dataset also
supports the fact that crime has been steadily
decreasing across the city as a whole. Due to
the fact that this data set does not include
shootings in its count of incidents, we
decided to analyze this specific incident in a
separate visualization due to its severity and
in order to compare it to overall crime in the
city.
4. Visualization III - Shootings
Figure 5 – Top Five Districts by Shootings,
2015-2017
4.1 Data
The dataset we used included information on
shootings reported in the city, spanning back
to 2015. We filtered on the top five districts
with the highest number of shootings
reported.
6
4.2 Results
Districts 22, 24, 25, 35 and 39 reported
having the most number of shootings. The
total number of shootings in these five
districts actually increased from about 600
in 2015 to over 650 in 2016.
4.3 Interpretations
Districts 24 and 25 were on both the top five
districts for overall crime and specifically
for shootings. This indicates that these
districts are particularly troubled in regard to
their crime rate. This also means that there
appears to be no direct correlation between
overall crime and shootings, since the other
three districts included in this visualization
do not appear on the previous one. The
findings we analyzed here also indicate that
gun violence is remarkably high in the city
of Philadelphia. In 2016 alone, there were
over 1,300 shootings. These top five districts
accounted for more than half of those
incidents.
This is also representative of the deadly
trend of gun violence across the United
States and confirms our original suspicion
that Philadelphia is no different. Although
the steadily decreasing crime rate is
promising, gun violence remains a major
issue for Philadelphia residents.
5. Visualization IV – Crime over
Population
Figure 6 – Crime in Districts 24 & 25 with Population
5.1 Data
This visualization was produced using the
same data set as the first. Here we looked at
the count of crime incidents over the course
of 2016. Since 2017 is not over yet, we
believed looking at the last full year of data
would be the best. We only looked at the
Police Districts 24 & 25 since there were the
two which had the most crime and
7
shootings. This data was overlaid over the
population by city block using the data set
built in to Tableau.
5.2 Results
The visualization showed there were not
many crimes committed in the middle, blue
section of the two districts. This blue
section indicates a higher population of
people live in that area. Crime incidents are
pretty spread out throughout the districts as
there are many smaller dots throughout the
visualization. While it is spread out, we can
see a concentration of incidents in the lower,
middle of the visualization. In this area,
there are multiple larger dots which
represent a greater number of crime
happened in that specific area.
5.3 Interpretations
In previous visualizations, we could see that
there is a numerous amount of crime which
takes place in Police Districts 24 & 25.
Here we can see the crime is not necessarily
concentrated in a specific area. There is one
area which has multiple location along the
street with multiple incidents. Kensington
Avenue is the street which multiple crime
occurs. This is backed up by news reports
of a lot of drugs and prostitution on this
street. [4]
6. Conclusion
Through visualizing crime data from
Philadelphia, we can see there is a log of
crime throughout the city. Most crime is
situated in the North Philadelphia area. This
area is within the Philadelphia Police
Department’s Districts 24 & 25. These two
districts are in the top five for both crime
overall and shootings. We were able to plot
all the crime data to see most crime is seen
in the area around Kensington Avenue.
News article support this data as well from
talking to people along Kensington Avenue
about drugs and prostitution in the area. The
data supports that this area is one of the
most crime filled area of Philadelphia.
Overall the crime is on a downward trend in
the city. Over the past few years the data
shows that crime and shootings are
decreasing in number of incidents.
While there was a lot of data contained in
the datasets from the Philadelphia Police
Department, there is room to improve the
dataset. The major issue with the data sets is
while the crime dataset has shootings, the
shooting dataset has additional columns.
These columns contain information specific
to shootings. These two datasets are
difficult to connect due to differences in data
and lack of a unique identifier similar in
each dataset. The major difference in the
data is with the crime codes. The crime
codes are more specific in the shooting
dataset then the complete crime dataset. An
additional issue is with the data is many
crimes categorized under “All Other
Offenses.” This type of crime was the
largest and without any way to understand
what these offences are creates difficulties in
analyzing the data. If these data issues were
fixed additional analysis can be completed
to get an improved picture of the crime in
Philadelphia.
8
References:
1. Uniform Crime Reporting Statistics, Federal Bureau of Investigation, 26 Jan. 2017,
www.ucrdatatool.gov/Search/Crime/Local/RunCrimeJurisbyJuris.cfm.
2. Palmer, Chris. “Police: Philly Crime at Lowest Level in Decades.” Philly.com, 11 Jan. 2017,
www.philly.com/philly/news/20170112_Police__Philly_crime_at_lowest_level_in_decades.html
3. Gramlich, John. “5 Facts about Crime in the U.S.” Pew Research Center, 21 Feb. 2017,
www.pewresearch.org/fact-tank/2017/02/21/5-facts-about-crime-in-the-u-s/.
4. Deeney, Jeff. “Philadelphia's Kensington Avenue: Heroin, Prostitution, and No Police.” The
Daily Beast, The Daily Beast Company, 13 Aug. 2011, www.thedailybeast.com/philadelphias-
kensington-avenue-heroin-prostitution-and-no-police.
5. “Crime Incidents.” OpenDataPhilly, City of Philadelphia,
www.opendataphilly.org/dataset/crime-incidents.
6. “Police Districts.” OpenDataPhilly, City of Philadelphia,
www.opendataphilly.org/dataset/police-districts.
7. “Police Stations.” OpenDataPhilly, City of Philadelphia,
www.opendataphilly.org/dataset/police-stations.
8. “Shooting Victims.” OpenDataPhilly, City of Philadelphia,
www.opendataphilly.org/dataset/shooting-victims.