HW 01: Gun Violence

Introduction

Gun violence in the US results in tens of thousands of deaths and injuries annually, and every year it seems as if the number of incidents involving guns increases. At the time of creating this lab, the US has experienced 268 mass shootings that fit the Mass Shooting Tracker project criterion in this year alone. Their criterion for a mass shooting is one in which four or more persons shot in one incident, at one location, at roughly the same time. We are going to examine some data from 2014-2017 about gun violence incidents in the US. The data are obtained from Kaggle, and have been slightly modified. We will also be using data about the U.S. Census and gun laws.

Getting started

Go to the course GitHub organization and locate your Lab 03 repo, which should be named hw-01-gun-violence-[GITHUB USERNAME]. Grab the URL of the repo, and clone it in RStudio. Don’t forget to configure git:

library(usethis)
usethis::use_git_config(user.name = "github username", user.email = "your email")

The variable descriptions are as follows:

incident_id: ID number
date: date of incident
state: state in which incident occurred
city_or_county: city or county in which incident occurred
address: address where incident occurred
n_killed: number of people killed
n_injured: number of people injured, but not killed
n_guns_involved: number of guns involved
doy: day of year (month/day) of incident
year: year
month: month

Exercise 1

In which five cities or counties did the most incidents occur? Answer the question using pipes (%>%) and the tidyverse functions we’ve discussed. Show your code and output and state your answer in a sentence.

Exercise 2

While there are a variety of definitions of mass shootings, one definition is “any incidents in which four or more people were shot, whether injured or killed”. Based on this definition, display the 10 incidents that meet this definition of mass shooting in descending order of number of people shot. Only display the date, state, city or county, and number of people.

Exercise 3

Let’s examine which days of the year are most common for gun violence. Display a table of the top six days of the year with highest number of total incidents, displayed in descending order. Comment on what you notice.

Exercise 4

Create a data frame called state_incidents where each row is an observation with a state and its total number of incidents.

Exercise 5

⊕Hint: Use the reorder() function while plotting.

⊕Hint: Use the stat = “identity” in your geom_() function. This lets us override the default option of just plotting the total counts.

Let’s look at the number of gun violence incidents by state. Plot a bar graph with the counts of gun violent incidents by each state. Display the graph such that the bars run horizontally, and also in descending order. Don’t forget to adjust the labels on your plot! Briefly comment on what you see.

Exercise 6

We expect that the more populated states would have a larger number of gun violence incidents. To address this, we want to add in some notion of population size. The dataset census.csv provides each state’s estimated population in 2017, as well as some other information related to that state. The variables are as follows:

state: state
state_code: two-letter postal code
population: estimated population in 2017
gun_laws: number of gun laws in 2017

First, let’s combine the two datasets together, such that for every observation in state_incidents, we add a column for corresponding population. Save this into state_incidents.

Exercise 7

Now let’s look at the number of gun violence incidents per 100000 residents in each state, and visualize it. First, create a new variable that is equal to the number of gun violence incidents per 100000 people. This is calculated as taking the number of incidents divided by number of residents, multiplied by 100000. Save this into the same data frame state_incidents, then provide a plot of these incidents per 100000 residents by state, once again displaying in descending order. How does this plot compare to that of Exercise 5?

Exercise 8

⊕Hint: Use the case_when() function.

Let’s create groupings for the state populations for easier visualization. Create a new variable called pop_group and group the states into the following categories: “small”, “medium”, “large”, “x-large” based on if their populations are less than 5 million, at least 5 million and less than 10 million, at least 10 million and less than 19 million, and 19 million or more. Save this into the same dataframe state_incidents.

Exercise 9

Another column in the census.csv dataset is the number of gun laws/provisions by state in 2017. Plot the incidents per 100000 residents by the number of gun laws in each state. Color each observation by its pop_group. Does there appear to be a relationship between number of gun laws and number of gun violence incidents per 100000 residents by state? What about between the number of gun laws and population size?