Mass casualty Incidents 2006 Onward

Intro

Our team decided to look at the data regarding mass casualty events. The various steps of the assignment helped us to narrow down the best way to proceed. Sketching allowed us to explore which categories of the data would be displayed and in which manner. The readings of this module were especially helpful in guiding us through these new principles. The first video explaining Altair was comprehensive and including most of what we need to know to begin. Some of the more impactful articles are cited in the following paragraphs.

Figure 1. Death count specified into category of affliction

Data

Our dataset comes from research carried out by the Associated Press, USA Today, and Northeastern University. The data tracks incidents of mass casualty in the United States from 2006 — present day, and is active with the last update to the data occurring on February 12th, 2025. An incident of mass killings is defined in the data as the intentional killing of four or more victims within a 24-hour period, excluding any deaths from the offender(s) or the unborn. The definition does not discriminate on method (ex shootings, arson), location, victim-stranger relationship, or number of locations over the 24 hour period.

Researchers gathered this data by flagging potential incidents from the Federal Bureau of Investigation’s Supplementary Homicide Report (SHR). These flagged incidents were verified using local law enforcement reports, court documents, and academic journal articles.

Design Process

To get started on our design process, our first step began with understanding what the data was showing us in order to successfully craft what we wish to relay to our audience. This involved viewing the data outside of the csv module to gain a clear understanding of what our data is demonstrating and how we can use that to craft a story. We created very basic visualizations using Altair of different variables throughout our data set. With these basic visuals, we were able to gain an understanding of what variables we wanted to focus on and move into the process of sketching.

Figure 2. Design process visualized

The variables chosen to focus on during the design process included…

• Date of Incident
• City
• Number of Victims Killed
• Number of Victims Injured
• Primary Method of Harm (Weapon Used)
• Location Type
• If an Assault Rifle was Used

Sketches

In order to decide what narrative we sought for this data to tell, we first explored a wide range of visualization techniques and types of visuals to utilize. This process ensured that our group got to explore the different mechanisms in which the chosen variables could be presented without honing in on one particular type of visual.

“Visualization tools can help you understand the data”

Figure 3 cont. Our sketches of possible designs

The benefits of sketching may best be demonstrated in the writing by Giorgia Lupi: Sketching With Data Opens The Mind’s Eye where she writes “ Generally speaking, drawing plays an important role in the production and communication of knowledge, and in the formation of new ideas” (Lupi, 2016).

Get Luke Lommatzsch’s stories in your inbox

Join Medium for free to get updates from this writer.

Remember me for faster sign in

During the ideation process, we explored possibilities for misleading visuals, using tactics such as y-axis distortion and cherry picking data to create false and misleading narratives surrounding the frequency of mass casualty incidents.

Figure 3 cont. Our sketches of possible designs

With our wide range of sketches, we sought help from our peers and discussed amongst each other to come to a consensus on the best way to represent the data visually. Qualitative peer feedback (from those outside the class) showed an inclination towards the simpler models, such as the line graphs depicting the frequency of mass casualty incidents over the years. Peers noted that these were most effective in depicting the story being told, whereas others (notably the sketch depicting weapon types with number of shootings) seemed overwhelming and confusing to read. We took this feedback, also keeping in mind that many of the critiques could be overcome with different design strategies during the creation process in Altair.

Figure 3 cont. Our sketches of possible designs

Prototypes

This step explored how our sketching and ideations could be converted into true visualizations in Altair.

Figure 4. Pie chart displaying first cause of death

Through the use of what we learned from homework one we were able to create scatter, bar, and pie charts efficiently. The pie charts we felt were especially useful in displaying the proportions of attacks done as stated in the reading 39 studies about human perception in 30 minutes “Not only did he find that pie charts were read as easily as, quickly, and accurately as bar charts, but that as the number of components in the chart increased, bars become less efficient encoding the data. The opposite was true for pie charts” (Elliott, 2019). The data we were working with seemed best suited for display in these manners. The cause of death felt like the most flexible to work with which lead to the creation of these graphs

Figure 5. Bar chart displaying location frequency

Convergence

The stacked bar chart was a clear front runner for its ability to display the large amount of data we were attempting to visualize. The brar chart shows which category of firearm is used most often in these attacks over the years. The other scatter plot was another choice for its ability to display which cause of death has the highest occurrence. Therefore, these four types of graphs we chose would represent the data we were looking to display in the best way possible.

Figure 6. Scatter plot displaying cause of death

Figure 7. Stacked bar chart displaying type of firearm