Visualizing US police shooting

Lab Reports


Police shooting has become one of the most concerned topics for the U.S. society recently. People are worried about the system being unfair and thus elicit unstable sentiments from the general public. Therefore, I decided to examine information related to the shooting crime for the recent 5 years to answer the question: Are there problems with police shooting, and if yes, what are they?

By visualizing the data, we could conclude there are some communalities of the police shooting. This is essential for our society to recognize the key points from this massive data so that the society could make some progress for the issues with police shooting.


Just like there are no two snowflakes that are identical, there are barely two cases that are identical. When it comes to single case of police shooting, there are a lot of things need to be considered. However, through viewing the Chicago Crime dataset, which includes the incidents of crime since 2001, I realized that by combining and analyzing the single factors, there are communalities that can be found to answer the big questions. For instance, “How does temperature affect the incident rate of violent crime (assault or battery)?” The data can point out the most inflammatory temperature for the police department and give them a sense of when and where to deploy more police forces in order to prevent more crimes. In this way, data can help the society detect problems and become safer.


The dataset being used for the lab is the US Police Shootings, which contains 4895 lines of records of police shootings from Jan, 2, 2015 to Jun, 15, 2020, with information on id, name, age, gender, race, date, manner of death, city, state, signs of mental illness, threat level, army category, flee method, and body cameras. The dataset was first imported to OpenRefine for cleaning. Meanwhile, the dataset was claimed to be cleaned already by Ahsen Nazir, the uploader of the dataset. Therefore, only two minor errors were discovered within the age column. The first one is displaying 37 with 37.1179 and the second one is inputting 6.0 instead of 60.0. After the two problems were fixed, the data was exported again as a csv file and further imported to Tableau Public.

With Tableau Public, I was able to create 9 charts: 3 pie charts, 1 tree map, 1 bubble chart, 1 bar chart, 1 density map, 1 area chart, and 1 scatter plot. I also used Tableau to combine the charts into workbooks and selected the same font and color for consistent expression.


I choose to present the charts in a way that the general public, such as newspaper readers, can easily read and understand at the first sight. I have divided these charts into four sections, so that people could looking at the shooting crime from four different aspects. The first section is the map, the overview of the number of shootings across states. The second and third sections are pertaining to the demographics and details of the people shot by the police. The last section is illustrating whether the police are wearing body cameras.


This is a US map that shows how many police shooting cases in each state. By merely looking at the map, people can find out the state with the most cases are California, Texas, and Florida. However, it is still too early to conclude that there are problems within these states, because there are many other factors to consider, such as the population within each state. The severity of the high number of shooting cases might be averaged out by the density of population.


Demographics is the most fundamental sections among the four sections. There are three charts in this section: Gender, Race, and Age. From these three charts, we can vividly see more males and whites are involved in the police shooting. Meanwhile, people aged at 37 were being shot the most times.


The section summarizes the conditions when police shooting took place. Factors include the type of weapon people held, the kind of interaction the people had with police, whether they have mental illness and whether they tried to flee from the police. By looking at these charts, we could see that most people being shot were with guns and attacked the police, but were without mental illness and did not tried to flee.


It is still hard to say for the above three sections if there are problems with police shooting, because people aged at 37 might also be the mode age that people commit crimes. Moreover, whether the person flee or have mental illness seems less crucial compared to the fact that most people being shot attacked the police. However, the fourth section represent the section that problem actually exists and awaits for change. The problem is that most police were not wearing body cameras while shooting happened even though wearing body cameras is necessary for many reasons, such as verification afterwards.


Two problems were uncovered through peer review. Firstly, my peer reported Alaska and Hawaii were hard to find on the world map and suggested me to move Alaska and Hawaii near to the American Continent for users to not omit them. Secondly, my peer reported the area map is surprising that there seemed to be more whites died from police shooting rather than other races, which is quite different from what people usually see from news. Again, there can be many reasons behind the scene. For example, there might be more whites among the whole population than the sum of the other races. Hence, I learned that my chart alone can be misleading in representing information. More research is needed before making any conclusions. In the future, more dataset should be examined and presented aside the existing graphs to give a more comprehensive and more complete result to the viewers.