Determine key attributes that lead to fatal police encounters.
Test initial hypotheses.
Create recommendations to help save lives on both sides.
Across the United States, there are altercations that result in killings by police officers, whether on-duty or off-duty. Each case is considered as a person dying from being shot, beaten, restrained, intentionally hit by a police vehicle, pepper-sprayed, tasered, or otherwise harmed by police.
For our capstone, we are addressing the issue of police violence that has become a heated topic throughout the country in recent years. As a team, we will be attempting to help law enforcement agencies spot risk factors so they may step in before risk transforms into actual harm.
The dataset we acquired from Mapping Police Violence has gathered over 9,000 police killings from January 2013 to May 2021. Analyzing the features surrounding fatal encounters will shed light on changes to be made to save more lives. After creating a classification model to predict the threat level of the victim, we've cast a larger net and acquired the 30,000+ Fatal Encounters dataset that was originally used to create the Mapping Police Violence dataset used in the first iteration. The final classification model we have created beat the baseline model by 20%.