Fatal Police Shootings in the United States
March 26, 2022
Updated March 26, 2022
Overview #
The Washington Post has spent the last several years collecting details of fatal shootings by on-duty police officers. Since 2015, over 7,000 cases have been added to this dataset.
For more details, check out the Washington Post article.
I intend to tregularly return to review and revise the analysis on this page as the underlying data is updated.
Why? #
I consider the work that the Washington Post has done on this topic to be extremely critical.
A systematic, data-driven analysis allows us to identify trends and patterns that might be missed by case-by-case reviews. It could inform pro-active action that could be undertaken to reduce the occurrences of fatal shootings.
I think it’s fair to say that any fatal shooting in a police encounter is a tragedy.
Data #
The underlying data can be accessed at the Washington Post’s Fatal Force Github repository.
Specifically, the raw data in .csv form can be downloaded here.
Here’s a glimpse of the full data as of March 26, 2022.
## Rows: 7,213
## Columns: 17
## $ id <dbl> 3, 4, 5, 8, 9, 11, 13, 15, 16, 17, 19, 21, 22,…
## $ name <chr> "Tim Elliot", "Lewis Lee Lembke", "John Paul Q…
## $ date <date> 2015-01-02, 2015-01-02, 2015-01-03, 2015-01-0…
## $ manner_of_death <chr> "shot", "shot", "shot and Tasered", "shot", "s…
## $ armed <chr> "gun", "gun", "unarmed", "toy weapon", "nail g…
## $ age <dbl> 53, 47, 23, 32, 39, 18, 22, 35, 34, 47, 25, 31…
## $ gender <chr> "M", "M", "M", "M", "M", "M", "M", "M", "F", "…
## $ race <chr> "A", "W", "H", "W", "H", "W", "H", "W", "W", "…
## $ city <chr> "Shelton", "Aloha", "Wichita", "San Francisco"…
## $ state <chr> "WA", "OR", "KS", "CA", "CO", "OK", "AZ", "KS"…
## $ signs_of_mental_illness <lgl> TRUE, FALSE, FALSE, TRUE, FALSE, FALSE, FALSE,…
## $ threat_level <chr> "attack", "attack", "other", "attack", "attack…
## $ flee <chr> "Not fleeing", "Not fleeing", "Not fleeing", "…
## $ body_camera <lgl> FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALS…
## $ longitude <dbl> -123.122, -122.892, -97.281, -122.422, -104.69…
## $ latitude <dbl> 47.247, 45.487, 37.695, 37.763, 40.384, 35.877…
## $ is_geocoding_exact <lgl> TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE…
Location #
The data includes latitude and longitude details for all the fatal shootings, which naturally lends itself to mapping.
Mental Illness #
One of the fields tracks whether or not the shooting victim was experiencing a mental health crisis at the time of the shooting.
It seems like there’s an extraordinarily high occurrence of mental health events at the time of the fatal shootings.
Solutions to handling scenarios manifested from mental health events might mitigate the occurrence of fatal shootings.
When we look at the proportion of fatal shootings by race that involved mental health events, we find that Asian, “Other”, or white had a much higher proportion ( greater than 20%) exhibiting signs of exhibiting mental illness.
One way to interpret this is that these racial categories were less likely to be shot in the absence of abnormal behavior.
On the other hand, victims identified as Black, Hispanic, or Native American were more likely to be shot even in the absence of signs of mental health issues.
A similar interpretation could be applied to fatal shootings based on a breakdown by gender.
Armed #
The armed field in the data tracks whether or not the victim was armed, and if so, with what sort of implement.
Alarmingly, there was a non-trivial number of unarmed victims in the dataset.
The armed field also tracks cases where the victim was in possession of a toy weapon.
Body Camera #
We have details about whether or not a body camera was in use.
A good number of fatal shooting events occurred when police body cameras were not in use.
Parting Thoughts #
If there’s any other way you’d like to see this data analyzed, feel free to drop me a note either in the comments below or through the contact form.