Shooting Events in Philadelphia, PA (2015 to 2023)

In [55]:

import carto2gpd
from folium.plugins import HeatMap, MarkerCluster
import folium
import geopandas as gpd
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import os
import osmnx as ox

Using OpenDataPhilly's API, I accessed the data on shooting events.

In [63]:

# Get shooting data from Open Data Philly's API
url = "https://phl.carto.com/api/v2/sql"
table_name = "shootings"
df = carto2gpd.get(url, table_name)

df = df.loc[df['year'] != 2023] # Use only full year's worth of data

df

Out[63]:

	geometry	cartodb_id	objectid	year	dc_key	code	date_	time	race	sex	...	offender_injured	offender_deceased	location	latino	point_x	point_y	dist	inside	outside	fatal
0	POINT (-75.24507 39.96569)	1	8772806	2022	202219019227.0	401	2022-04-22T00:00:00Z	19:36:00	B	M	...	N	N	200 BLOCK N FELTON ST	0.0	-75.245070	39.965688	19	0.0	1.0	0.0
1	POINT (-75.24559 39.97038)	2	8772807	2022	202219019249.0	401	2022-04-22T00:00:00Z	22:31:00	B	M	...	N	N	6300 BLOCK GIRARD AVE	0.0	-75.245588	39.970376	19	0.0	1.0	0.0
2	POINT (-75.24276 39.98001)	3	8772808	2022	202219019654.0	401	2022-04-25T00:00:00Z	08:05:00	B	M	...	N	N	6100 BLOCK W OXFORD ST	0.0	-75.242756	39.980014	19	0.0	1.0	0.0
3	POINT (-75.23637 39.96480)	4	8772809	2022	202219020571.0	401	2022-04-30T00:00:00Z	18:26:00	B	M	...	N	N	200 BLOCK N 58TH ST	0.0	-75.236365	39.964798	19	1.0	0.0	0.0
4	POINT (-75.23284 39.96565)	5	8772810	2022	202219023257.0	411	2022-05-15T00:00:00Z	21:29:00	B	M	...	N	N	5600 BLOCK VINE ST	0.0	-75.232836	39.965655	19	0.0	1.0	0.0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
14404	POINT (-75.16850 39.99893)	14405	8771815	2021	202139050068.0	111	2021-09-18T00:00:00Z	12:44:00	B	M	...	N	N	2200 BLOCK W CAMBRIA ST	0.0	-75.168500	39.998927	39	0.0	1.0	1.0
14405	POINT (-75.16381 39.99836)	14406	8771816	2021	202139050228.0	111	2021-09-19T00:00:00Z	11:12:00	B	M	...	N	N	2000 BLOCK W CAMBRIA ST	0.0	-75.163811	39.998360	39	0.0	1.0	1.0
14406	POINT (-75.15433 40.01967)	14407	8771817	2021	202139050716.0	116	2021-09-21T00:00:00Z	17:14:00	B	M	...	N	N	1700 BLOCK SAINT PAUL ST	0.0	-75.154326	40.019668	39	0.0	1.0	1.0
14407	POINT (-75.16315 40.00838)	14408	8771818	2021	202139051064.0	111	2021-09-23T00:00:00Z	20:07:00	B	M	...	N	N	3500 BLOCK N 21ST ST	0.0	-75.163154	40.008377	39	0.0	1.0	1.0
14408	POINT (-75.18036 40.00967)	14409	8771819	2021	202139051442.0	401	2021-09-25T00:00:00Z	18:30:00	B	F	...	N	N	3100 BLOCK HENRY AVE	0.0	-75.180359	40.009666	39	0.0	1.0	0.0

13683 rows × 23 columns

The line plot below shows a steady amount of shooting events from 2015 to 2019, with a sporadic increase in shooting events from 2019 to 2020.

In [29]:

# Organize counts of shootings by year
by_year = df.groupby('year').size().reset_index(name='Count')

# Set up the plot and plot the line graph
sns.set_style("darkgrid")
plt.figure(figsize=(10, 6))
sns.lineplot(x='year', y='Count', data=by_year, marker='o', color='#5eccab', linewidth=2.5)
plt.xlabel('Year')
plt.ylabel('Count')
plt.title('Count of Shootings - Philadelphia, PA (2015 to 2023)')
plt.gca().set_facecolor('black')
plt.tick_params(colors='black')
plt.xlabel('Year', color='black')
plt.ylabel('Count', color='black')
plt.show()

Using Folium, I plotted all of the data points, colored by the year they occurred. I leveraged marker clustering so that way not all ~14,000 points appeared on the map at once, but rather you can zoom into an area of interest. Using a tooltip, I added information about the year, age, sex, and race of the offender, if the offender was injured or deceased, and whether an officer was involved.

In [62]:

import folium
from folium.plugins import MarkerCluster

# Filter out rows with missing geometries
df2 = df.dropna(subset=['point_x', 'point_y'])

# Create a map centered on the calculated location
f = folium.Map(location=[39.99, -75.13], zoom_start=10, tiles='Cartodb dark_matter')

# Define a colormap for different years
cm = {
    2015: '#546319',
    2016: '#c0affb',
    2017: '#e6a176',
    2018: '#00678a',
    2019: '#984464',
    2020: '#5eccab',
    2021: '#cdcdcd',
    2022: '#ddcc77'
}

# Create a marker cluster to help with map performance and aesthetics
marker_cluster = MarkerCluster().add_to(f)

# Iterate over the filtered dataframe rows
for index, row in df2.iterrows():
    latitude = row['point_y']
    longitude = row['point_x']
    year = row['year']
    color = cm.get(year, year)
    race = row['race']
    sex = row['sex']
    age = row['age']
    officer_involved = row['officer_involved']
    offender_injured = row['offender_injured']
    offender_deceased = row['offender_deceased']
    tooltip = f"Year: {year}<br>Race: {race}<br>Age: {age}<br>Sex: {sex}<br>Officer Involved: {officer_involved}<br>Offender Injured: {offender_injured}<br>Offender Deceased: {offender_deceased}"
    
    folium.CircleMarker(
        location=[latitude, longitude],
        radius=5,
        alpha=0.5,
        color=color,
        fill=True,
        fill_color=color,
        tooltip=tooltip
    ).add_to(marker_cluster)

# Add title to the map
title_html = '''
             <h3 align="center" style="font-size:20px"><b>Shootings - Philadelphia, PA (2015 - 2022)</b></h3>
             '''
f.get_root().html.add_child(folium.Element(title_html))

# Fit the map bounds to include all the markers
f.fit_bounds(marker_cluster.get_bounds())

# Display the map
f

Out[62]:

Make this Notebook Trusted to load map: File -> Trust Notebook

I decided as another visualization, I wanted to make a heat map of fatal shootings. Heat maps are another great way of visualizing density of events. I filtered the data to only include fatal shooting events, and then used Folium again to create the heat map.

In [35]:

where = f"fatal = '1'"
fatal = carto2gpd.get(url, table_name, where=where)

In [20]:

fatal['point_y'] = fatal.point_y
fatal['point_x'] = fatal.point_x
coords = fatal[['point_y', 'point_x']]

# eliminate rows with missing values
coords = coords.dropna()

In [64]:

m = folium.Map(
    location=[39.99, -75.13],
    tiles='Cartodb dark_matter',
    zoom_start=11)

title_html = '''
             <h3 align="center" style="font-size:20px"><b>Fatal Shootings - Philadelphia, PA (2015 - 2023)</b></h3>
             '''
m.get_root().html.add_child(folium.Element(title_html))


HeatMap(coords).add_to(m)

m

Out[64]:

Make this Notebook Trusted to load map: File -> Trust Notebook