Informational Report on the US Police Shootings Dataset in Power BI
“Continue to speak out against all forms of injustice to yourselves and others, and you will set a mighty example for your children and for future generations.” — Bernice King
In 2020, the #BlackLivesMatter movement caught my attention, and I extensively researched the campaign through documentaries and articles. Questions like “Are these xenophobic crimes?” or “Is a specific race being targeted?” kept coming to me as I watched video after video and read paper after paper. I wondered if there was an objective perspective and data to support these events. Who was recording them all this while? These questions led me to discover a dataset on US Police shootings, and as a data analyst, I eagerly sought to uncover perspectives on this movement backed by data. The dataset I found on Kaggle contained records of US police shootings from 2015 to 2020.
It is important to note that the dataset comprises recorded cases, and various factors could have affected the data collection process. In this post, I will share how I explored the data in Power BI and present my informational report without additional analysis or personal sentiments. Let’s dig in!
An informational report contains factual information and seeks to inform the audience without additional analysis.
The Data
The dataset consists of 4894 rows, each reflecting a unique death, with 4673 male victims and 222 female victims. Although this information contextualizes the data, it is essential to note that it was discovered during the investigation. I loaded the data into Power BI Desktop and identified questions based on columns that piqued my curiosity and had the potential to tell a story:
NB: The only cleaning I had to do was binning the numerical column, Age. Check out my post on Binning Numerical columns without writing DAX.
The questions I wrote down included:
- What is the scope of the data? Which years do the shootings cover?
- Where in the US were these cases recorded?
- What are the ages of the victims?
- What is the gender distribution of the victims?
- What are the racial backgrounds of the victims?
- Did the police officers have body cameras during the cases?
- How did the victims die? What is the manner of death?
- Did the victims pose any level of threat?
- Did the victims try to flee?
- Were the victims carrying any weapons, and if so, what kind?
To harmonize the visualization, I divided it into two sections: Demographics and Manner of Death.
Visualisations
Demographics
States
Out of the total shootings, 701 occurred in California, 426 in Texas, 324 in Florida, and 222 in Arizona. The other states had less than 200 recorded cases.
Ages
Among the victims, 1549 were in their 30s, 1370 were in their 20s, 883 were in their 40s, 585 were in their 50s, 289 were teenagers, and 236 were senior citizens. (For the criteria used to bin the ages, please refer to my post on Binning.)
Race
Of the victims, 2475 were White, 1298 were Black, 902 were Hispanic, 93 were Asian, and 78 were Native.
Manner of Death
This section explores how the victims died and the circumstances surrounding the law enforcement officers and the victims at the time of the shootings.
Presence of Police Body Camera
In 4316 cases, the police officers did not have body cameras (the presence of body cameras was false). However, in the other 578 cases, body cameras were present.
Manner of Death
Out of the total victims, 4646 were shot, and 248 of them were shot and tasered.
Level of Threat
The levels of threat were categorized into Attacked, Other and Undetermined. 3160 of the victims Attacked, 1527 fell in the Other Category, and 207 of the levels were Undetermined.
Possession of Arms
Of the victims, 2764 had guns, 818 had sharp objects, 418 carried Unknown weapons, and 347 were Unarmed.
Tried Fleeing
3072 were Not Fleeing, 820 tried Fleeing in cars, 642 on Foot, and 360 used Other means.
Shootings Across the Years
From 2015 to 2020, the number of recorded cases per year was as follows: 965 in 2015, 903 in 2016, 906 in 2017, 888 in 2018, 858 in 2019, and 374 in 2020.
Throughout this post, I shared my thought process while visualizing the US police shooting dataset and presented my informational report on the visuals. I also provided a brief definition of an informational report. If you had the same data and visuals, how would you analyze this information?
PS: check out my first designs for the same report :<
