## Introduction

Gun violence results in thousands of deaths and injuries in the USA annually.According to data collected by the Centers for Disease Control and Prevention, there were 73505 non fatal firearm injuries,11208 homicides,21175 suicides,505 deaths due to accidental/negligent discharge of a firearm; and 281 deaths due to firearms-use with “undetermined intent” in 2013 alone.

In 2010, 67% of homicides in the USA involved the use of firearms.Gun violence cost U.S taxpayers approximately $516 Million in direct hospital bills.Poor urban and rural areas which are associated with gang violence,are the epicenters, often involving male juveniles and young male adults.

## Injured and Killed

## Time Series Plot

From the plot above we see that the number of injuries increases when the number of individuals killed increases.The number of injured individuals were high during the start of September, end of October and the beginning of November.

## Month Wise

Most of the injuries and killings took place in the month of October followed by November and September.

## Day Wise

According to the data a large portion of the deaths and injuries related to gun violence took place on Sundays. The number of such incidents reduced through Monday and the rest of the week, taking a dip on Wednesdays. The number of incidents increased from Wednesdays onwards.

## Relationship Between Killed and Injured

## Linear Plot and Correlation

The correlation between the number of teenagers killed and teenagers injured is about 0.919. A strong correlation implies a strong linear relationship.

## Linear Model

According to the model that has statistically significant variables,for every 1000 increase in teenagers injured due to gun violence, the number of teenagers killed would increase by 730. Let’s try to evaluate this linear model using some checks.

## Q-Q Plot

This Q-Q plot shows that the residuals are normally distributed.A normally distributed residuals plot is essential in the strength of a linear model.

## Residual histogram plot

The histogram plot is nearly normal. We need to have a normal distribution for a strong linear model.

## Residual Plot

The residual plots are not randomly distributed. For a linear model to be effective, these points have to be randomly distributed around the zero line.

## Numbers killed by city

Chicago leads the country in terms of number of teenagers killed in the three months due to gun violence. Chicago has one of the highest murder rates in the country.The 2015 year-end crime statistics showed there were 468 murders in Chicago in 2015 compared with 416 the year before, a 12.5% increase, as well as 2900 shootings, 13% more than the previous year and up to 29% since 2013.

As per Illinois state laws, purchasing a handgun has a 72 hour waiting period after the sale is made. The waiting period for a rifle/shotgun is 24 hours.Chicago formerly prohibited the sale of firearms within city limits, but on January 6,2014, a federal judge ruled that this was unconstituitional.

## Further Exploration

We can explore the data further by mapping the crimes to a county level US map. This would allow us to discern any underlying patterns.

## Source

https://en.wikipedia.org/wiki/Gun_violence_in_the_United_States

https://en.wikipedia.org/wiki/Chicago#Crime

https://en.wikipedia.org/wiki/Gun_laws_in_Illinois

You can find the code here