Introduction
There have been many detrimental side effects of the Covid-19 Pandemic within NYC this past year aside from the tragic loss of so many to the virus. A recent New York Times article reported that crime rates during 2020, specifically shootings, have surged and reached rates that haven’t been seen in roughly a decade. From the article: “Experts say the economic and physical strain of the virus, which disproportionately took lives and jobs from neighborhoods that were already struggling with high levels of gun violence, most likely drove the rise in shootings.”
As a longtime NYC resident, I was interested in seeing if the NYPD’s records would demonstrate this rapid increase of shootings as reported by the Times. I also wanted to look into which boroughs these incidents were occurring in, as well as the gender and age profiles of the perpetrators and victims.
Exploring the Data
The dataset that I used for my review was NYC Open Data’s record of shooting incidents from 2006-2020 as compiled by the NYPD. Each row within the set provided information about the specific time and location of the shooting, as well as demographic information about the perpetrators and victims.
In order to make a deeper investigation into the dataset I used Tableau Public, a free platform to explore, create and publicly share data visualizations online. Once the dataset was uploaded to the platform I was able to isolate and filter the various quantitative and categorical dimensions in order to create visualizations, look for trends and extract more information from the data. I was inspired by a similar exploration of shooting occurrences that was done on a national scale as shown in the visualizations below:
Visualizations and Findings
The first visualization I created was to see if the total number of shooting incidents in 2020 matched the numbers from a decade ago. Using Tableau, I was able to break down the total number of shootings into their respective years to create a line graph that charted the previous steady decline and current surge. From the graph you can clearly see that, after a decade of predominantly decreasing numbers, the amount of shooting incidents from 2020 has indeed increased to where it was 10 years prior.
Breaking the total number of shootings down by hour of occurrence enabled me to compare the times of day when they were happening. As I expected, more shootings occur over night than during the day. I found that the below area chart effectively communicated the stark variance between when the majority of shootings occur. The hours around 9AM show the lowest incidents (177) and the highest (1994) happen around 23 hours military time/11PM.
After exploring the time of day when these incidents were happening the next logical step was to consider the boroughs where they were happening. The following text table lists the total number of shootings over the years in each borough. The corresponding packed bubble visualization works to emphasize the difference in the levels of occurrence between each borough. To further highlight this scale, I used a color gradient, where the highest numbers (Brooklyn) had the darkest hue and the lowest occurrences (Staten Island) had the lightest hue.
After determining the when and the where, my final investigation into this dataset was to look at who was involved in these shootings. I first looked at the gender breakdown of who was committing these shootings. It is worth noting that the dataset did not account for alternative genders such as trans or non-binary so the results are confined to male and female. After filtering out incidents where the gender was unknown or not listed so as to not skew the results, I was able to break down the total shootings into male and female perpetrators. While I had expected a higher percentage of male perpetrators, I was surprised by just how many more male perpetrated shootings there were (13,305) versus female (334). Males also constituted the vast majority of the shooting victims (90.68%), with the conclusion being that the bulk of shooting incidents involve male on male interactions.
Another aspect that I looked into was the age of the shooters as well as the victims. After filtering out any instances where the age was unknown, I was able to see that the majority of shooting perpetrators fell into the 18-24 year old group, compared to the majority of victims who were older and fell into the 25-44 year old group. For both victims and perpetrators, the 65+ age group saw the smallest numbers. In these charts as well as previous visualizations, I tried to avoid using abrasive colors such as dark reds or blacks. My goal was not to trivialize the content but rather to emphasize the findings versus the dark nature of the subject matter.
Future Investigations
While this dataset can provide some limited insight on who is involved in these shootings, as well as the when and where, it does not explain why they are happening. I purposely chose not to explore the racial profiles of the perps/victims, with the understanding that such an analysis should be much more nuanced. Similar to the limitations created by only allowing two genders, reducing a person to the few, broad race categories listed would not yield the full picture of why these shootings are occurring. Further investigation should compare the socio-economic profiles of the precincts/neighborhoods where the higher incidents of shootings are occurring, as well as of the perpetrators and shooters themselves. There are so many outside factors that contribute to criminal activity outside of age, gender or race. Such factors should be addressed and investigated in order to ascertain a more in depth understanding of the implications of this dataset.