Project Status: Completed
Jupyter Notebook Viewer
Read Article

• Project Objective
• Methods Used
• Technologies
• Project Description
• Project Results
• Installation

(Back to top)
The purpose of this project is to examine the factors that play into the horrible event of a fatal shooting by the police in the US. Which ones carry more weight that lead to fatal shootings and are perhaps predictive in nature? Are they race? State location? Mental Illness? Based on our findings from the dataset of the available variables, we are looking to predict the deceased’s race or mental illness status.

(Back to top)

• Data Cleaning
• Exploratory Data Analysis
• Data Visualization
• Machine Learning

(Back to top)

• Python
• Pandas
• NumPy
• re
• Seaborn
• Matplotlib
• Copy
• Geopandas
• Folium
• Geopy
• Nominatim
• Template
• MacroElement
• PrettyTable
• Sklearn - preprocessing
• Sklearn.linear_model - LogisticRegression
• Sklearn.preprocessing - StandardScaler
• Sklearn.metrics - accuracy_score
• Sklearn.svm - SVC
• Sklearn.tree - DecisionTreeClassifier
• Sklearn.ensemble - RandomForestClassifier
• Sklearn.model_selection - cross_val_score
• Sklearn.model_selection - RandomizedSearchCV

(Back to top)

• A dataset from the Washington Post was used, which had over 5700 data points and were collected between 2015-2020.
• Cleaned the data by using pandas.
• As far as feature engineering, 9 out of the total 17 variables type had to be transformed into different types.
• To read more about the data cleaning process click here.

(Back to top)
With race being a big question going into this project, seeing the number of victims based on race seemed to be perfectly logical. Here we see that within all of the dataset White was killed the most and then black was almost half of the white percentage.

Following now in the same line of questioning, we took a look at the physical location by state with respect to the races that were shot in those states:

Next, we wanted to see if shootings might have some kind of relation based on state.

In the plot shown below, we can see that the distribution of gender, race and signs of mental illness are unbalanced.

In the next plot we pay our attention to the feature "body cam footage" and check how that might determine whether an individual was shot. Below we can see that manny officers did have the camera off making this data skewed and hard to have as a determining variable.

Below we can see a box plot that represents the age distribution by race. The age distribution by race has a strong representation across all between the ages of 22 and 47, with a mean of 37. Please note, we replaced the 262 null values in the “age” column with the mean values based on race and gender.

In order to create a model to predict signs of Mental Illness, we used the following ML algorithms:

• Logistic Regression
• SVC (Support Vector Classification)
• SGD (Stochastic Gradient Descent)
• Decision Tree
• Random Forest We train our model and get the following accuracy scores.

• bootstrap: [True,False],
• max_depth: [int(x) for x in np.linspace(start = 10, stop = 110, num =11)],
• max_features: [“auto”,”sqrt”],
• min_samples_split: [2,5,10],
• min_samples_leaf: [1,2,4],
• n_estimators: [int(x) for x in np.linspace(start = 200, stop = 2000, num =10)]

Our model improved from 0.724 to 0.772. Next we implement our model to the test set and we get an accuracy score of 0.771, which is the percentage of correctly predicted labels.

In regard to predicting the race, we concluded that we would need to gather further data in order to create an acceptable model to predict the race. For more details click here.

Please click here for final conclusion.

(Back to top)

• Clone this repo (for help see this tutorial).
• Raw data, data processing/transformation script is being kept in this repo. Click here for notebook.
• Note: If GitHub doesn't load the notebook please refer to Jupyter Notebook Viewer.