Data Study on Mass Shootings in the US

Posted on Feb 16, 2016
The skills the author demoed here can be learned through taking Data Science with Machine Learning bootcamp with NYC Data Science Academy.
Contributed by Matt Samelson. He  is currently in the NYC Data Science Academy 12 week full time Data Science Bootcamp program taking place between January 11th to April 1st, 2016. This post is based on his first class project - R visualization, due on the second week of the program.

Mass Shootings Are High Profile But Should Not Be Central To The Gun Debate

Gun issues in the United States are contentious and polarizing.  Passion runs high for both proponents and opponents of gun control.  In a social climate in which the consequences for misrepresentation and egregious distortion are few, individuals on both sides speak freely on principal and not from holistic data evaluation.

Gun event data is sparse and the quality limited.  There has been no government-sponsored research (a-la CDC, etc.) for decades.  This is largely due to politics.  That said, there is crowd-sourced data on the internet.  There is no denying that gun related deaths and injuries occur daily in the United States.

Mother Jones Magazine, a liberally-oriented investigative journalism publication, has published numerous articles on "mass shootings."   There is no universally accepted definition of a Mass Shooting in the gun debate.  Generally speaking,  Mass Shootings involve a high level of victims.  Mother Jones defines a Mass Shooting as an event with 4 or more fatalities.

Mother Jones Magazine was challenged by activists in response to assertions made in several articles.  In response, the magazine made it's data publicly available.

This analysis focuses on understand and attempting to elicit statistical significance from the Mother Jones data.

Data on Mass Shooting Events

The geography of US mass shootings between 1982 and 2015 is illustrated in the map below.

The map highlights the 73 events contained comprising the Mother Jones data.  The size of the event pertains to number of total victims.

Dragging your cursor over an event will provide event summary information.

The incidence of a shooting and number of victims have no relation to geography.

The following chart's highlight the leading states in terms of events and victims:

Data Study on Mass Shootings in the US

California leads the Nation in both mass shooting incidents and victims with 12 out of 73 events and approximately 200 victims.  Florida and Washington follow distantly in terms of incidents with 6 each.  Texas and Colorado follow in terms of victims with 110 each.

Data on Victims:  Fatalities and Wounded

Any act of random violence is an atrocity.  However, news media and interest groups often play on details to imply trends and characteristics of shooting events that just aren't there.  While most outlets attempt to shock us with the magnitude of those wounded or killed in a particular event, big numbers are not consistent across all (or even a majority) of events.  See the figures below.


Data Study on Mass Shootings in the US


Data Study on Mass Shootings in the US


Data Study on Mass Shootings in the US

30 events - just under half of total events - had less then total victims.  Just over 50 events had between 5 and 10 fatalities while just over 40 events had 0 to 5 injured.  Overall, events with large numbers of wounded and fatalities were indeed a minority of events.  While tragic and high profile - they just didn't happen all that often.

Perpetrator Characteristics

Perpetrator sex, age, and race are shown in the figures below.


Mass shooting perpetrators are most often male, most often white, and are most often 40 to 45 years old.  It is critically important to realize that these characteristics are independent.  In other words, the hypothetical "typical" perpetrator (if there is such a thing) is not white and male and aged 40 to 45.

Perpetrators are distinctly male.  The figure above reveals males conduct incidents in stunning proportion.  In fact, 70 of 73 incidents were perpetrated by single male actors (one event was performed by a male/female team and two events by lone female assailants).

A sizable proportion of perpetrators are white.  46 of 73 events were perpetrated by actors who were white.  This is still a standout figure.  Whites were trailed by blacks who conducted 12 events and by Asians with 6 events.  Latinos, Native Americans and Other ethnicities each accounted for under 5 events.

Male Perpetrators

Males are the actors in virtually all events.  Yet how does race and age fit into the equation?  See the figures below.


While males between the ages of 40 and 45 are the largest group of perpetrators, white males constitute only about half of the group.  The figure above illustrates that white males, by virtue of their number, are heavily represented in most age groups.

White Male Perpetrators

A density plot of white males conditioned on age reveals the likelihood that a male of a particular ethnicity falls in a particular age group.



While there are many white male perpetrators in the 40 to 45 year age range, white male perpetrators are more likely to be in their early 20s.


Density functions by race conditioned upon age suggest that differences exist in age among perpetrators of different race. The density function for white males, illustrated earlier and shown in the figure above in pink, is readily visible with a peak for individuals in their early 20s and a secondary peak for males in their early 40s.  Black men who perpetrate these shootings are most often in their late 30s.  Other races (Latino, Native American, Asian, and "other") also tend to most often be in their late 30s.

The densities show the likelihood of a perpetrator of a particular race being of a certain age.  While it provides insight into historical data, it is not enough to assert that race and age are statistically associated.

Mosaic Plot

A mosaic plot with a Chi Squared test of observed vs expected values provides a formalized, rigorous statistical assessment as to the relationship of race and age among male perpetrators.  In short, a mosaic plot enables us to visually compare categorical variables (race and age group are categorical variables - age itself is a continuous) and determine statistically if our expectations for the group combinations is in line with the counts we actually observe.

The mosaic plot below illustrates our data in slightly amended groups (grouping the data as shown was necessary for a meaningful test because of a small sample size).  Grey boxes indicate that observations are in line with expectations and that knowing age will not provide insight into whether a perpetrator is white (and vice versa).  Blue or red colored boxes indicate deviations from expectations.   Blue indicates greater than expected values and red less than expected values.   In short, a blue or red box for one or more of the groups would indicate that knowing the age of a perpetrator would provide insight into that person's race or vice versa.


The grey coloring of all boxes and the insignificant Chi-Squared p-value of .3114 shown under the legend tell us that we are unable to gain insight into perpetrator race knowing age group (i.e., the variables are independent).

Data on Weapon Legality and Mental Illness

Perpetrators obtained weapons legally and displayed "irregular behavior" prior to incidents in many events.



The preponderance of an individual to obtain legal weapons and display unusual behavior in close to 60 percent of event.   There is certainly much to debate around restrictions concerning the legal purchase of guns.  There is nothing in the data that permits meaningful elaboration on this issue.

The fact that others saw "irregular" behavior prior to the shooting in nearly 60 percent of cases is another matter.  Erratic behavior that may result in harm to others is something that can readily be acted upon.  It is usually pretty clear when individuals act in depressed or despondent manners.  Most people can tell that "something is wrong".

Still, most people fail to take action on an individual that may cause harm to themselves or others.  Part of the reluctance may be fear of being wrong or involving themselves in someone else's business.  I think the failure to act in these cases is selfish and cowardly.  There is little doubt in my mind that at least some of these incidents could have been avoided if someone who "saw something" said something to law enforcement or mental health experts.

The Worst Events

I have purposefully avoided mentioning specific events until this point.  My intention was to have readers focus on events from a statistical and analytical standpoint as opposed to becoming blinded by sensational elements of particular events.  That said, no discussion of mass shootings would be complete without a list of the most egregious events.


3 0f the 5 events are quite recent, occurring since 2009.  All events dominated national media coverage when they occurred and some continue to do so.  It is notable that these events are all unusual even for mass shootings.  Note the fatality, wounded, and victims numbers and compare with the histograms above.

It is clear that the number of casualties far exceed those of the preponderance of events.  The point is that shootings such as these are a genuine rarity.  While they highlight some of the most atrocious conduct one can imagine they are extreme events.  They are not indicative of daily gun related events that occur in the United States nor are they common for their class of events.  Media and activists focusing on these events and implying they are commonplace and central to gun control and rights issues are incorrect, uninformed and likely to be shifting emphasis on these events for their own particular objectives.


Mass shooting are atrocious.  No question about it.  However, scientifically evaluating the Mother Jones Mass Shooting Data yields the following insights:

  • There have been only 73 events in which 4 or more individuals have been killed since 1982.
  • The most prevalent events have fewer than 10 victims and few than 10 fatalities.
  • Total victims constitute an extreme minority of total gun incident victims.
  • Weapons in most events were obtained legally.
  • Most perpetrators displayed "unusual" or "erratic" or "disturbing" behavior prior to their event.
  • Mass shootings with substantial media coverage tend to be those with high victim count.  These are generally outlying events within the mass shooting data set.

In regard to perpetrators:

  • Majority are male.
  • Most are aged between 40 and 45 years old.
  • Most are white.
  • We are unable to gain any statistically-based predictive insight to associate age and race within the male sex category.

Mass shootings, while terrible, constitute a small proportion of gun related events.  The worse the event, the greater the media scrutiny and the hyped the coverage.  Furthermore, most events involve legally obtained weapons and prior indications of erratic behavior.  Mass shootings should certainly an area of focus for legal and medical professionals.  However, the data simply does not constitute the incorporation of mass shooting data in the broader discussion of gun rights/gun control debate.  Political, media, and activist incorporation simply takes events out of context, "muddies the waters" and makes meaningful gun-related discussions more complicated.


About Author

Matt Samelson

Matt Samelson is a data scientist and leader passionate about "hands-on" problem-solving using statistical analysis, predictive analytics, and visualization. He has a track record of driving incremental business improvements and a background in management, consulting, and quantitative research....
View all posts by Matt Samelson >

Related Articles

Leave a Comment

No comments found.

View Posts by Categories

Our Recent Popular Posts

View Posts by Tags

#python #trainwithnycdsa 2019 2020 Revenue 3-points agriculture air quality airbnb airline alcohol Alex Baransky algorithm alumni Alumni Interview Alumni Reviews Alumni Spotlight alumni story Alumnus ames dataset ames housing dataset apartment rent API Application artist aws bank loans beautiful soup Best Bootcamp Best Data Science 2019 Best Data Science Bootcamp Best Data Science Bootcamp 2020 Best Ranked Big Data Book Launch Book-Signing bootcamp Bootcamp Alumni Bootcamp Prep boston safety Bundles cake recipe California Cancer Research capstone car price Career Career Day citibike classic cars classpass clustering Coding Course Demo Course Report covid 19 credit credit card crime frequency crops D3.js data data analysis Data Analyst data analytics data for tripadvisor reviews data science Data Science Academy Data Science Bootcamp Data science jobs Data Science Reviews Data Scientist Data Scientist Jobs data visualization database Deep Learning Demo Day Discount disney dplyr drug data e-commerce economy employee employee burnout employer networking environment feature engineering Finance Financial Data Science fitness studio Flask flight delay gbm Get Hired ggplot2 googleVis H20 Hadoop hallmark holiday movie happiness healthcare frauds higgs boson Hiring hiring partner events Hiring Partners hotels housing housing data housing predictions housing price hy-vee Income Industry Experts Injuries Instructor Blog Instructor Interview insurance italki Job Job Placement Jobs Jon Krohn JP Morgan Chase Kaggle Kickstarter las vegas airport lasso regression Lead Data Scienctist Lead Data Scientist leaflet league linear regression Logistic Regression machine learning Maps market matplotlib Medical Research Meet the team meetup methal health miami beach movie music Napoli NBA netflix Networking neural network Neural networks New Courses NHL nlp NYC NYC Data Science nyc data science academy NYC Open Data nyc property NYCDSA NYCDSA Alumni Online Online Bootcamp Online Training Open Data painter pandas Part-time performance phoenix pollutants Portfolio Development precision measurement prediction Prework Programming public safety PwC python Python Data Analysis python machine learning python scrapy python web scraping python webscraping Python Workshop R R Data Analysis R language R Programming R Shiny r studio R Visualization R Workshop R-bloggers random forest Ranking recommendation recommendation system regression Remote remote data science bootcamp Scrapy scrapy visualization seaborn seafood type Selenium sentiment analysis sentiment classification Shiny Shiny Dashboard Spark Special Special Summer Sports statistics streaming Student Interview Student Showcase SVM Switchup Tableau teachers team team performance TensorFlow Testimonial tf-idf Top Data Science Bootcamp Top manufacturing companies Transfers tweets twitter videos visualization wallstreet wallstreetbets web scraping Weekend Course What to expect whiskey whiskeyadvocate wildfire word cloud word2vec XGBoost yelp youtube trending ZORI