Jovian
Sign In
Learn practical skills, build real-world projects, and advance your career

Exploratory Data Analysis on Police Violence In United States

Analysing the risk of being killed by police use of force in United States by Age, Race-Ethnicity and Sex

alt

Preface

The killing George Floyd by a white police officer in Minneapolis shook the nation and set off massive protests around the world in 2020. The police system in US is in question. In this analysis, I wanted to analyze both the sides - fatal encounters in US and Police killings in US.

The analysis is truly based on the data I've used and on the articles I've researched as part of the analysis. Please refer the links.

According to Human Rights Watch reports, Police in the United States kill nearly 1,000 people every year. Black and Indigenous people were approximately three times more likely than white people to be fatally shot by police. 98.3 percent of killings by police from 2013- 2020 have not resulted in officers being charged with a crime. Between 2005 and 2015, only 54 officers were charged after police-involved killings.

As per BBC News, there were 14,400 gun-related homicides, 23,941 suicides involving guns, 38,355 total deaths involving guns in 2019. US ranks first among the top 10 civilian gun-owning contries followed by Yemen, Serbia, Monetenego, Uruguay, Canada.

After declining in late 1990s, U.S. gun suicide and gun murder rates have edged higher in recent years as per Pew Research Center.

Police can use deadly force if they merely perceive a threat, and not whether the shooting victim actually posed a threat according to Vox.

Black Americans are killed at a much higher rate than White Americans. Most victims are young male between 20 and 40 years old. See more details: Washington Post.

It is not just the civilians who are killed. More details: Officer Down Memorial Page. We'll also look at the Police Officer's who are killed in the line of duty.

In our analysis, we'll analyze the data to come up with insights to prove the above facts.

Let's begin!

Introduction

In this notebook, we'll use a real-world dataset from Kaggle that contains Police Violence & Racial data in United States.

The kaggle dataset files will be downloaded using opendatasets helper library.

About the Real-World Data

For the analysis, we will use the Police Violence & Racial Equity - Part 1 of 3 data by JohnM on Kaggle.

The data is available in 3 parts and we have used Part 1 of the data. The contains the below informations.
Part 1: Citizen deaths, police deaths, and other outcomes

  • Police shootings
  • Citizen fatalities involving police
  • Police officer deaths suffered in the line of duty

Part 1 contains five .csv datasets.

  1. deaths_arrests.csv
  2. fatal_encounters_dot_org.csv
  3. police_deaths_538.csv
  4. police_killings_MPV.csv
  5. shootings_wash_post.csv

We will be considering ONLY the below datasets for our analysis.

  • fatal_encounters_dot_org.csv
  • police_deaths_538.csv

Let us analyze each of these datasets in detail.

fatal_encounters_dot_org.csv
  • The dataset is updated as of 8/2020
  • The dataset has 10 columns and 28,621 rows
  • Columns that we consider for the analysis are:
    • 'Unique ID',
    • "Subject's age",
    • "Subject's gender",
    • "Subject's race",
    • 'Date of injury resulting in death (month/day/year)',
    • 'Location of death (state)',
    • 'Location of death (county)',
    • 'Latitude',
    • 'Longitude',
    • 'Agency responsible for death',
    • 'Cause of death',
    • 'Intentional Use of Force (Developing)',
    • 'Date (Year)'
police_deaths_538.csv
  • This file has the Names and data for police killed in the line of duty.
  • The dataset is updated as of 8/2020
  • The dataset has 29 columns and 22,801 rows
  • Columns that we consider for the analysis are:
    • 'dept'
    • 'cause_short'
    • 'date'
    • 'year'
    • 'state'

Let's see an outline of the steps we will follow for the analysis:

  1. Download the dataset using the opendatasets Python library
  2. Perform data preparation and cleaning
  3. Perform exploratory analysis and visualization
  4. Ask and answer questions about the data
  5. Summarize inferences and write a conclusion
  6. Future Work
  7. References
jovian.commit()