RoboCops

Analysis of Fatal Police Shootings

This project aims to comprehensively analyze fatal police shootings in the United States, delving into the underlying factors and motivations driving these tragic events.

Authors

RoboCops - Peter, Sachin, Arun,Surajit, Christian, Valerie

School of Information, University of Arizona

Introduction

The police shooting data set is sourced from a Kaggle dataset originally created by Karolina Wullum. The data, consisting of shootings for all fifty states and DC, was originally logged by The Washington Post after the Black Lives Matter movement was born from the shooting of Michael Brown in 2014. The dataset contains information on fatal shootings that occurred from 2015 to 2017 and highlights information on fatal cases as well as the demographics of the city the individuals were shot in. The goal of this project is to contextualize police shootings to further understand how and why these unfortunate situations occur.

Why the Police Killings dataset?

Police shootings are a sad reality of policing practices that back decades. While many are justified to protect the lives of innocent civilians, there has been an uptick in recent years of police shooting people whom they perceive to be a threat in the heat of the moment, but are proven innocent after the fact. These shootings can often be hard to analyze because there are many psychological factors that come into play when an officer decides to discharge their weapon, such as their training, the perceived threat, and an officer’s internalized prejudice. Without further information on the psyche of a police officer, the reason why they discharge their weapon cannot be determined. However, their actions and surroundings can be analyzed to provide insight into how and why a shooting occurs.

Analysis Goals

Our analysis aims to understand the dataset on police killings, extending its scope beyond the original focus tracked by the Washington Post. We plan to include socio-economic factors like housing income, poverty rates, and population demographics. While initially associated with the Black Lives Matter movement, our goal is to explore how socio-economic factors intersect with incidents of police violence.

By incorporating these additional factors, we hope to uncover correlations and patterns that provide insight into the broader societal context surrounding police use of force. Through data visualization and analysis, we aim to contribute to a deeper understanding of these issues and support evidence-based policy-making for promoting equity and justice.

Dataset

Main Dataset

# Load in the datasets
fatality <- read.csv("data/PoliceKillingsUS.csv", na.strings = "")
median_income <- read.csv("data/MedianHouseholdIncome2015.csv")
poverty_perc <- read.csv("data/PercentagePeopleBelowPovertyLevel.csv")
hs_perc <- read.csv("data/PercentOver25CompletedHighSchool.csv")
race_perc <- read.csv("data/ShareRaceByCity.csv")
city_pop <- read.csv("data/us-cities-top-1k-multi-year.csv")
state_pop <- read.csv("data/Population_Estimate_data_Statewise_2010-2023.csv")

The PoliceKilllingsUS.csv is a dataset sourced from Kaggle using the following link:

Fatal Police Shootings in the US

The .csv contains data between 2015 and 2017 and was compiled by The Washington Post. The dataset is comprised of 14 variables and 2535 shootings. An in depth description of the variables can be seen in the following table:

Variable	Description	Data Type	Values
id	Data ID	numeric	2535 entries
name	Name of deceased	character	Names
date	Date of shooting	date	dd/mm/yy from 2015-2017
manner_of_death	Means of death	categorical	shot, shot and tasered
armed	Weapon/Tool	categorical	69 categories
age	Age of deceased	numeric	6-91 years old
gender	Gender of deceased	logical	F/M
race	Race of deceased	categorical	A, B, H, N, O, W, Unknown
city	City	categorical	1417 cities
state	State	categorical	State 2 letter abbreviations and DC
signs_of_mental_illness	Sign of mental illness during incident	logical	True/False
threat_level	Deceased's threat level	categorical	attack, other, undetermined
flee	Method to evade police	categorical	Car, Foot, Not fleeing, Other
body_camera	Reports of police with body camera	logical	True/False

Data Cleaning

According to another Kaggle repository on the same topic and sourcing the same resources (https://www.kaggle.com/datasets/mrmorj/data-police-shootings), race codes unknown values as ““. To avoid confusion, the values are recoded as Unknown using the mutate function.

Additionally, a secondary dataframe was generated by grouping the dataset by city and race and acquiring the counts. This will allow multiple avenues of analyses such as ascertaining whether or not the proportions of shootings by race in each city differs from the proportions of race in the cities in the census datasets.

City Census Datasets

In addition to the dataset containing the fatal police shooting in the repository, there are four additional 2015 census datasets each containing over 29000 observations:

MedianHouseholdIncome2015.csv
PercentagePeopleBelowPovertyLevel.csv
PercentOver25CompletedHighSchool.csv
ShareRaceByCity.csv

Data	Description	Number of Cities
PercentOver25CompletedHighSchool.csv	Contains the percentage of individuals 25 years and older that are highschool graduates for cities	29329
MedianHouseholdIncome2015.csv	Contains the median income for cities	29322
PercentagePeopleBelowPovertyLevel.csv	Contains the percentage of the population below poverty for cities	29329
ShareRaceByCity.csv	Contains the demographics for cities	29268

Each census dataset contains two common columns: Geographic.Area and City. The values for City is name of the city and a classifier (ie Alexander City city).

In addition to these two columns, each dataset contains information regarding the city’s demographic information: percent of highschool graduates, median income, poverty rate, and proportion of races (White, Black, Native American, Asian, Hispanic). Each of these values are numeric.

Data Cleaning

For ease of analysis, the four census datasets are concatenated together by city into one merged dataset.

Additionally, the City value is edited to remove the classifier from each label so that the city value between the main dataset and the census datasets will be equivalent. Doing so will ease analysis.

City Population Data

In addition to city level census data, city populations were collected from the plotly dataset repository (https://github.com/plotly/datasets/blob/master/us-cities-top-1k-multi-year.csv). The .csv contains populations of the 1000 largest US cities from 2014 to 2018. The dataset contains 4000 observations and six variables detailed below:

Variable	Description	Data Type	Values
City	City	character	925 entries
State	State	character	50 States and District of Colombia
Population	Population	numeric	10048-8405837
lat	Latitude	numeric	21.31-61.22
lon	Longitude	numeric	-157.86 to -70.26
year	Year	numeric	2014-2018

State Population Data

In addition to city level population data, a statewise population dataset was collected. Population_Estimate_data_Statewise_2010-2023.csv contains population estimates from 2010-2023 for all fifty states, the District of Colombia, and Puerto Rico. The dataset has 15 variables including state and populations estimates for each year from 2010-2023.

Questions

How do socioeconomic factors such as median income, poverty rates, and educational relate with incidents of police use of force across different geographic areas?
To what extent do racial demographics, including the proportions of White, Black, Hispanic, Asian, and Native American residents, correlate with incidents of police use of force within communities?

Analysis plan

Question 1

How do socioeconomic factors such as median income, poverty rates, and educational relate with incidents of police use of force across different geographic areas?

Approach

In this part of our analysis, we aim to investigate whether police use of force is influenced by fundamental societal inequalities. This investigation is divided into two parts. Firstly, we will explore the relationship between the severity of police use of force and the level of poverty in communities. To do this, we will utilize the PercentagePeopleBelowPovertyLevel.csv dataset, which provides insights into poverty rates across different areas. Secondly, we will examine any potential correlation between socio-economic status, particularly income & educational levels, and incidents of police killings. This approach will offer valuable insights into the broader relationship between socio-economic factors and police use of force.

Visualizations

Plot 1

To visually represent our findings, we propose using bar plots or column plots. These plots will display the distribution of threat levels perceived by police officers, categorized by poverty rate brackets or educational attainment levels. Annotations will be included where applicable to provide additional context and highlight significant observations.

Plot 2

For the second plot, we will employ regression analysis and half-eye plots for regressed variables. This approach will allow us to visualize the relationship between income levels and incidents of police killings more comprehensively. This plot will be compared to plot 1 to provide a clearer understanding of any potential correlations between socio-economic factors and police use of force.

Interpretation of Plots

Overall, these visualizations will help us gain insights into the complex interplay between socio-economic factors and incidents of police use of force. By exploring these relationships, we aim to contribute to a deeper understanding of the underlying societal dynamics driving police interactions and identify potential avenues for addressing systemic inequalities.

Datasets and variables to be used

Primary Dataset (Fatal Police Shooting)

Used Variables

manner_of_death: Manner of death of victims (shot or tasered)

threat_level: The level of threat the victim posed (Attack, Other, Undetermined)

Supplementary Datasets

PercentPeopleBelowPovertyLevel.csv: he median income for cities.

ShareRaceByCity: Contains the demographics for cities

Variables to be created

To be determined as the analysis progresses

Question 2

To what extent do racial demographics, including the proportions of White, Black, Hispanic, Asian, and Native American residents, correlate with incidents of police use of force within communities?

Approach

For this segment of our analysis, we aim to investigate whether police have a higher propensity to use force against specific racial groups. This investigation is divided into two parts to comprehensively understand the dynamics at play. Firstly, we will examine the severity of police actions relative to population changes over time. This analysis will be overlaid with racial population composition to discern any patterns or disparities. In the second part, we will explore whether the use of specific instruments by police varies across racial groups and age demographics. In the absence of specific data, we will prorate the racial population distribution even for incidents of police killings to facilitate comprehension.

Visualizations

Plot 1

To visually represent our findings, we propose using leaflet or ggmap plots alongside column plots. These plots will introduce a new variable, Shooting Intensity, calculated as the ratio of Police Shootings to Population. Further, we will zoom into select major states, incorporating racial distribution as an additional factor.

Plot 2

For the second plot, we will utilize heatmaps and stacked bar plots to illustrate the impact of specific instruments concerning Shooting Intensity across racial groups and age distributions. This approach will provide a comprehensive visualization of the relationship between racial and age demographics in police killings.

Interpretation of Plots

Overall, these visualizations will provide valuable insights into whether police use of force is driven by deliberate choices or influenced by inherent biases or chance factors related to crime.

Datasets and Variables to be used

Primary Dataset (Fatal Police Shootings):

Used variables

race: The race of the victim of the police shooting

manner_of_death: The manner in which victims lost their lives (shot, tasered, etc)

city: The city or community where the shooting occured

Supplementary Datasets

ShareRaceByCity: Contains the demographics for cities

Variables to be created

shooting Intensity: Police shootings / Population

Timeline

Task Name	Status	Assignee	Due	Priority	Summary
Proposal-For Peer Review	Complete	All	April 1	High	Write proposal for peer review
Peer Review	Complete	All	April 1	High	Write up peer reviews for other groups
Proposal-Updated	Complete	All	April 8	Medium	Update proposal using peer feedback
Proposal-Final	Complete	All	April 15	Low	Update proposal using instructor feedback
Question 1: Bar Plots	Complete	Surajit	April 22	Medium	Generate plot/slides for plot
Question 1: Density and Regression Plots	Complete	Peter	April 22	Medium	Generate plot/slides for plot
Question 1: Interpret	Complete	Christian	April 22	Medium	Interpret plots
Question 2: Shooting Intensity 1	Complete	Arun	April 22	Medium	Generate plot/slides for plot
Question 2: Shooting Intensity 2	Complete	Sachin	April 22	Medium	Generate plot/slides for plot
Question 2: Interpretation	Complete	Valerie	April 22	Medium	Interpret plots
Presentation Structure	Complete	Christian		Medium	Create presentation order
Write-up Abstract/Introduction	Complete	Christian		Medium	Write up introduction and abstract

Disclaimer

There is no evident ethical concern because the data is publically available, but in consideration of the sensitivity of the topic, names of the deceased will not be included for the analysis.

Potential inherent biases in the dataset include classification of police shooting, the fact that the dataset only consists of fatal shootings, and breadth of reported incidents. While the inherent dataset does not appear to have variables to address these biases, these concerns can be addressed in future analyses.

References

https://www.kaggle.com/datasets/kwullum/fatal-police-shootings-in-the-us/data

https://github.com/plotly/datasets/blob/master/us-cities-top-1k-multi-year.csv

--- title: "RoboCops" subtitle: "Analysis of Fatal Police Shootings" author: - name: "RoboCops - Peter, Sachin, Arun,Surajit, Christian, Valerie" affiliations: - name: "School of Information, University of Arizona" description: "This project aims to comprehensively analyze fatal police shootings in the United States, delving into the underlying factors and motivations driving these tragic events." format: html: code-tools: true code-overflow: wrap code-line-numbers: true embed-resources: true editor: visual code-annotations: hover execute: warning: false --- ```{r} #| label: load-pkgs #| message: false #| echo: false #| output: false pacman::p_load(kableExtra, plotly) ``` ## Introduction The police shooting data set is sourced from a Kaggle dataset originally created by Karolina Wullum. The data, consisting of shootings for all fifty states and DC, was originally logged by The Washington Post after the Black Lives Matter movement was born from the shooting of Michael Brown in 2014. The dataset contains information on fatal shootings that occurred from 2015 to 2017 and highlights information on fatal cases as well as the demographics of the city the individuals were shot in. The goal of this project is to contextualize police shootings to further understand how and why these unfortunate situations occur. ## Why the Police Killings dataset? Police shootings are a sad reality of policing practices that back decades. While many are justified to protect the lives of innocent civilians, there has been an uptick in recent years of police shooting people whom they perceive to be a threat in the heat of the moment, but are proven innocent after the fact. These shootings can often be hard to analyze because there are many psychological factors that come into play when an officer decides to discharge their weapon, such as their training, the perceived threat, and an officer's internalized prejudice. Without further information on the psyche of a police officer, the reason why they discharge their weapon cannot be determined. However, their actions and surroundings can be analyzed to provide insight into how and why a shooting occurs. ## Analysis Goals Our analysis aims to understand the dataset on police killings, extending its scope beyond the original focus tracked by the Washington Post. We plan to include socio-economic factors like housing income, poverty rates, and population demographics. While initially associated with the Black Lives Matter movement, our goal is to explore how socio-economic factors intersect with incidents of police violence. By incorporating these additional factors, we hope to uncover correlations and patterns that provide insight into the broader societal context surrounding police use of force. Through data visualization and analysis, we aim to contribute to a deeper understanding of these issues and support evidence-based policy-making for promoting equity and justice. ## Dataset ### Main Dataset ```{r} #| label: load-dataset #| message: false # Load in the datasets fatality <- read.csv("data/PoliceKillingsUS.csv", na.strings = "") median_income <- read.csv("data/MedianHouseholdIncome2015.csv") poverty_perc <- read.csv("data/PercentagePeopleBelowPovertyLevel.csv") hs_perc <- read.csv("data/PercentOver25CompletedHighSchool.csv") race_perc <- read.csv("data/ShareRaceByCity.csv") city_pop <- read.csv("data/us-cities-top-1k-multi-year.csv") state_pop <- read.csv("data/Population_Estimate_data_Statewise_2010-2023.csv") ``` The `PoliceKilllingsUS.csv` is a dataset sourced from Kaggle using the following link: <a href=https://www.kaggle.com/datasets/kwullum/fatal-police-shootings-in-the-us/data>Fatal Police Shootings in the US</a> The .csv contains data between 2015 and 2017 and was compiled by The Washington Post. The dataset is comprised of 14 variables and 2535 shootings. An in depth description of the variables can be seen in the following table: ```{r} #| label: dataset-table #| echo: false #| message: false # Create columns for data fatality_var <- colnames(fatality) fatality_desc <- c("Data ID", "Name of deceased", "Date of shooting", "Means of death", "Weapon/Tool", "Age of deceased", "Gender of deceased", "Race of deceased", "City", "State", "Sign of mental illness during incident", "Deceased's threat level", "Method to evade police", "Reports of police with body camera") fatality_type <- c("numeric", "character", "date", "categorical", "categorical", "numeric", "logical", "categorical", "categorical", "categorical", "logical", "categorical", "categorical", "logical") fatality_values <- c("2535 entries", "Names", "dd/mm/yy from 2015-2017", "shot, shot and tasered", "69 categories", "6-91 years old", "F/M", "A, B, H, N, O, W, Unknown", "1417 cities", "State 2 letter abbreviations and DC", "True/False", "attack, other, undetermined", "Car, Foot, Not fleeing, Other", "True/False") # Bind the vectors into a dataframe fatality_table <- as.data.frame(cbind(fatality_var, fatality_desc, fatality_type, fatality_values)) colnames(fatality_table) <- c("Variable", "Description", "Data Type", "Values") row.names(fatality_table) <- seq(1:14) # Generate a nice table with kable my_table <- kable(fatality_table, "html") %>% kable_styling(full_width = FALSE) # Display the table my_table ``` #### **Data Cleaning** According to another Kaggle repository on the same topic and sourcing the same resources (<a href=https://www.kaggle.com/datasets/mrmorj/data-police-shootings>https://www.kaggle.com/datasets/mrmorj/data-police-shootings</a>), race codes unknown values as "". To avoid confusion, the values are recoded as Unknown using the `mutate` function. Additionally, a secondary dataframe was generated by grouping the dataset by `city` and `race` and acquiring the counts. This will allow multiple avenues of analyses such as ascertaining whether or not the proportions of shootings by race in each city differs from the proportions of race in the cities in the census datasets. ### City Census Datasets In addition to the dataset containing the fatal police shooting in the repository, there are four additional 2015 census datasets each containing over 29000 observations: - `MedianHouseholdIncome2015.csv` - `PercentagePeopleBelowPovertyLevel.csv` - `PercentOver25CompletedHighSchool.csv` - `ShareRaceByCity.csv` ```{r} #| label: census-table #| echo: false #| message: false # Create columns for data census_data <- c("PercentOver25CompletedHighSchool.csv", "MedianHouseholdIncome2015.csv", "PercentagePeopleBelowPovertyLevel.csv", "ShareRaceByCity.csv") census_desc <- c("Contains the percentage of individuals 25 years and older that are highschool graduates for cities", "Contains the median income for cities", "Contains the percentage of the population below poverty for cities", "Contains the demographics for cities") census_n <- c(nrow(hs_perc), nrow(median_income), nrow(poverty_perc), nrow(race_perc)) # Bind the vectors into a dataframe census_table <- as.data.frame(cbind(census_data, census_desc, census_n)) colnames(census_table) <- c("Data", "Description", "Number of Cities") row.names(census_table) <- seq(1:4) # Generate a nice table with kable my_table <- kable(census_table, "html") %>% kable_styling(full_width = FALSE) # Display the table my_table ``` Each census dataset contains two common columns: `Geographic.Area` and `City`. The values for `City` is name of the city and a classifier (ie Alexander City city). In addition to these two columns, each dataset contains information regarding the city's demographic information: percent of highschool graduates, median income, poverty rate, and proportion of races (White, Black, Native American, Asian, Hispanic). Each of these values are numeric. #### Data Cleaning For ease of analysis, the four census datasets are concatenated together by city into one merged dataset. Additionally, the `City` value is edited to remove the classifier from each label so that the city value between the main dataset and the census datasets will be equivalent. Doing so will ease analysis. ### City Population Data In addition to city level census data, city populations were collected from the plotly dataset repository (<a href=https://github.com/plotly/datasets/blob/master/us-cities-top-1k-multi-year.csv>https://github.com/plotly/datasets/blob/master/us-cities-top-1k-multi-year.csv</a>). The .csv contains populations of the 1000 largest US cities from 2014 to 2018. The dataset contains 4000 observations and six variables detailed below: ```{r} #| label: city-table #| echo: false #| message: false # Create columns for data city_var <- colnames(city_pop) city_desc <- c("City", "State", "Population", "Latitude", "Longitude", "Year") city_type <- c("character", "character", "numeric", "numeric", "numeric", "numeric") city_values <- c("925 entries", "50 States and District of Colombia", "10048-8405837", "21.31-61.22", "-157.86 to -70.26", "2014-2018") # Bind the vectors into a dataframe city_table <- as.data.frame(cbind(city_var, city_desc, city_type, city_values)) colnames(city_table) <- c("Variable", "Description", "Data Type", "Values") row.names(city_table) <- seq(1:6) # Generate a nice table with kable my_table <- kable(city_table, "html") %>% kable_styling(full_width = FALSE) # Display the table my_table ``` ### State Population Data In addition to city level population data, a statewise population dataset was collected. `Population_Estimate_data_Statewise_2010-2023.csv` contains population estimates from 2010-2023 for all fifty states, the District of Colombia, and Puerto Rico. The dataset has 15 variables including state and populations estimates for each year from 2010-2023. ## Questions 1. How do socioeconomic factors such as median income, poverty rates, and educational relate with incidents of police use of force across different geographic areas? 2. To what extent do racial demographics, including the proportions of White, Black, Hispanic, Asian, and Native American residents, correlate with incidents of police use of force within communities? ## Analysis plan ### **Question 1** How do socioeconomic factors such as median income, poverty rates, and educational relate with incidents of police use of force across different geographic areas? #### **Approach** In this part of our analysis, we aim to investigate whether police use of force is influenced by fundamental societal inequalities. This investigation is divided into two parts. Firstly, we will explore the relationship between the severity of police use of force and the level of poverty in communities. To do this, we will utilize the `PercentagePeopleBelowPovertyLevel.csv` dataset, which provides insights into poverty rates across different areas. Secondly, we will examine any potential correlation between socio-economic status, particularly income & educational levels, and incidents of police killings. This approach will offer valuable insights into the broader relationship between socio-economic factors and police use of force. #### **Visualizations** ##### **Plot 1** To visually represent our findings, we propose using bar plots or column plots. These plots will display the distribution of threat levels perceived by police officers, categorized by poverty rate brackets or educational attainment levels. Annotations will be included where applicable to provide additional context and highlight significant observations. ##### **Plot 2** For the second plot, we will employ regression analysis and half-eye plots for regressed variables. This approach will allow us to visualize the relationship between income levels and incidents of police killings more comprehensively. This plot will be compared to plot 1 to provide a clearer understanding of any potential correlations between socio-economic factors and police use of force. ##### **Interpretation of Plots** Overall, these visualizations will help us gain insights into the complex interplay between socio-economic factors and incidents of police use of force. By exploring these relationships, we aim to contribute to a deeper understanding of the underlying societal dynamics driving police interactions and identify potential avenues for addressing systemic inequalities. ##### **Datasets and variables to be used** **Primary Dataset (Fatal Police Shooting)** **Used Variables** **manner_of_death:** Manner of death of victims (shot or tasered) **threat_level:** The level of threat the victim posed (Attack, Other, Undetermined) **Supplementary Datasets** **PercentPeopleBelowPovertyLevel.csv:** he median income for cities. **ShareRaceByCity:** Contains the demographics for cities **Variables to be created** To be determined as the analysis progresses ### **Question 2** To what extent do racial demographics, including the proportions of White, Black, Hispanic, Asian, and Native American residents, correlate with incidents of police use of force within communities? #### **Approach** For this segment of our analysis, we aim to investigate whether police have a higher propensity to use force against specific racial groups. This investigation is divided into two parts to comprehensively understand the dynamics at play. Firstly, we will examine the severity of police actions relative to population changes over time. This analysis will be overlaid with racial population composition to discern any patterns or disparities. In the second part, we will explore whether the use of specific instruments by police varies across racial groups and age demographics. In the absence of specific data, we will prorate the racial population distribution even for incidents of police killings to facilitate comprehension. #### **Visualizations** ##### **Plot 1** To visually represent our findings, we propose using leaflet or ggmap plots alongside column plots. These plots will introduce a new variable, `Shooting Intensity`, calculated as the ratio of Police Shootings to Population. Further, we will zoom into select major states, incorporating racial distribution as an additional factor. ##### **Plot 2** For the second plot, we will utilize heatmaps and stacked bar plots to illustrate the impact of specific instruments concerning `Shooting Intensity` across racial groups and age distributions. This approach will provide a comprehensive visualization of the relationship between racial and age demographics in police killings. ##### **Interpretation of Plots** Overall, these visualizations will provide valuable insights into whether police use of force is driven by deliberate choices or influenced by inherent biases or chance factors related to crime. ##### **Datasets and Variables to be used** **Primary Dataset (Fatal Police Shootings):** **Used variables** **race:** The race of the victim of the police shooting **manner_of_death:** The manner in which victims lost their lives (shot, tasered, etc) **city:** The city or community where the shooting occured **Supplementary Datasets** **ShareRaceByCity:** Contains the demographics for cities **Variables to be created** **shooting Intensity:** Police shootings / Population ## Timeline ```{r} #| label: timeline-table #| echo: false #| message: false # Create vectors to represent the tasks tasks <- c("Proposal-For Peer Review", "Peer Review", "Proposal-Updated", "Proposal-Final", "Question 1: Bar Plots", "Question 1: Density and Regression Plots", "Question 1: Interpret", "Question 2: Shooting Intensity 1", "Question 2: Shooting Intensity 2", "Question 2: Interpretation", "Presentation Structure", "Write-up Abstract/Introduction") # Create vector to represent status status <- c("Complete", "Complete", "Complete", "Complete", "Complete", "Complete", "Complete", "Complete", "Complete", "Complete", "Complete", "Complete") # Create vector to represent ownership member <- c("All", "All", "All", "All", "Surajit", "Peter", "Christian", "Arun", "Sachin", "Valerie", "Christian", "Christian") # Create vector to represent the deadline due <- c("April 1", "April 1", "April 8", "April 15", "April 22", "April 22", "April 22", "April 22", "April 22", "April 22", "", "") # Create vector to represent priority priority <- c("High", "High", "Medium", "Low", "Medium", "Medium", "Medium", "Medium", "Medium", "Medium", "Medium", "Medium") # Create vector to represent description description <- c("Write proposal for peer review", "Write up peer reviews for other groups", "Update proposal using peer feedback", "Update proposal using instructor feedback", "Generate plot/slides for plot", "Generate plot/slides for plot", "Interpret plots", "Generate plot/slides for plot", "Generate plot/slides for plot", "Interpret plots", "Create presentation order", "Write up introduction and abstract") # Bind the columns table_2 <- as.table(cbind(tasks, status, member, due, priority, description)) colnames(table_2) <- c("Task Name", "Status", "Assignee", "Due", "Priority", "Summary") row.names(table_2) <- seq(1:nrow(table_2)) # Generate a nice table with kable timeline_table <- kable(table_2, "html") %>% kable_styling(full_width = FALSE) # Display the table timeline_table ``` ## Disclaimer There is no evident ethical concern because the data is publically available, but in consideration of the sensitivity of the topic, names of the deceased will not be included for the analysis. Potential inherent biases in the dataset include classification of police shooting, the fact that the dataset only consists of fatal shootings, and breadth of reported incidents. While the inherent dataset does not appear to have variables to address these biases, these concerns can be addressed in future analyses. ## References <a href=https://www.kaggle.com/datasets/kwullum/fatal-police-shootings-in-the-us/data>https://www.kaggle.com/datasets/kwullum/fatal-police-shootings-in-the-us/data</a> <a href=https://github.com/plotly/datasets/blob/master/us-cities-top-1k-multi-year.csv>https://github.com/plotly/datasets/blob/master/us-cities-top-1k-multi-year.csv</a>