Developing Intimacy With Your Datasubject Police Killing

Developing Intimacy With Your Datasubject Police Killingsthis Exercis

Developing intimacy with your data subject: Police killings. This exercise involves working with existing datasets to examine, transform, and explore the data to develop a deep understanding of its properties and qualities. The datasets contain snapshot details of recorded deaths caused by US law enforcement agencies, sourced from The Guardian's "The Counted" and The Washington Post's "Fatal Force." For each dataset, you will analyze their meaning, physical properties, and differences, as well as undertake data cleaning, potential data augmentation, and visual exploration to gain insights and deepen your familiarity with the data's value.

Paper For Above instruction

Introduction

The analysis of datasets related to police killings in the United States provides critical insights into law enforcement practices, systemic issues, and societal impacts. This paper examines two prominent datasets sourced from The Guardian's "The Counted" and The Washington Post's "Fatal Force," focusing on understanding their data structures, properties, and insights through examination, transformation, and exploration phases. By thoroughly scrutinizing these datasets, we aim to identify differences, assess their representativeness, and explore ways to enhance data quality and analytical potential.

Examination of the Datasets

The datasets under discussion serve as snapshot records documenting fatalities caused by police operations across the US. The Guardian's "The Counted" dataset, compiled during 2015-2016, includes detailed information about each incident, such as demographics of the victims, the circumstances of the deaths, and contextual notes. It is structured in a tabular format with variables like date, location, victim demographics (age, gender, race), and incident details. The data type is mixed—dates are stored as date/time strings, categorical variables such as race and gender as text, and numerical variables like age as integers or floats. The total size of this dataset is approximately 987 entries, with consistent formatting and data completeness.

In contrast, The Washington Post's "Fatal Force" dataset extends from 2015 onwards and is more comprehensive, continuously updated with real-time entries. Its structure features similar variables but often includes additional fields such as law enforcement agency, mental health status, and whether the incident involved gunfire. The dataset is slightly larger, with over 8,000 entries. Physically, the data is stored in CSV format, with neatly organized columns, though some entries have missing data, especially in subjective categories like mental health status. The dataset’s size and scope suggest greater diversity in incident details and temporal coverage.

When comparing these datasets, both serve as valuable repositories but differ significantly in scope and depth. "The Counted" emphasizes human stories and demographic detail for a specific period, providing rich qualitative insights, while "Fatal Force" offers a broader, more extensive quantitative view suitable for longitudinal analysis. These differences impact their utility, with the Guardian data better suited for qualitative, demographic-focused studies, and the Washington Post data more apt for longitudinal trend analysis.

Transformation of the Data

Data cleaning is essential to ensure analytical accuracy. For "The Counted," this involves standardizing date formats, correcting inconsistencies in categorical variables like race and gender, and addressing missing values, perhaps via imputation or removal depending on their significance. For "Fatal Force," the focus should be on handling missing or incomplete incident details, validating location data, and harmonizing different variable formats for easier comparison.

Additional valuable data could include socioeconomic indicators of the incident locations, such as income levels, crime rates, or community health statistics. Incorporating such information would contextualize police killings within broader socio-economic frameworks, potentially revealing systemic correlations. Furthermore, integrating data on police policies, use-of-force incidents, and accountability measures could yield deeper insights into causative factors.

Transformation also entails normalization of data types, creating consistent coding schemes for categorical data, and establishing common variables if cross-dataset comparisons or aggregations are intended. For example, ensuring race categories align across datasets allows for direct comparisons in demographic analyses.

Exploration and Visual Analysis

Using tools like Excel, Tableau, or R, each dataset can be visually explored through various plots. In Excel, pivot tables and charts such as histograms or pie charts illustrate the frequency distribution of incidents across demographics and geographies. Tableau allows for interactive dashboards with maps indicating incident locations, and timelines highlighting trends over time.

In R, packages like ggplot2 and dplyr facilitate detailed visualizations. For instance, creating age distribution histograms reveals the prominence of certain age groups among victims. Mapping incident locations uncovers geographic hotspots of police killings. Trend lines over time highlight increases or decreases in incidents, which can be cross-referenced with policy changes or social events.

These visual explorations deepen understanding by revealing patterns, clusters, and anomalies—such as disproportionate casualties among specific racial groups or in certain regions. They also highlight data limitations, such as missing data or inconsistent recording practices, guiding future data collection efforts.

Conclusion

Analyzing these datasets on police killings illuminates the complexities of the data and its potential for critical insights into law enforcement practices. The examination clarified their structure, content, and differences, while transformation preparatory steps ensure data quality for analysis. Visual exploration provided tangible understandings of demographic, geographic, and temporal patterns. Together, these efforts foster nuanced perspectives on the systemic issues surrounding police-related deaths, emphasizing the importance of high-quality, comprehensive data in fostering accountability and social justice.

References

  • The Guardian. (2016). The Counted: People killed by police in the US. Retrieved from https://www.theguardian.com/us-news/ng-interactive/2016/jun/01/the-counted-police-killings-us-database
  • The Washington Post. (2023). Fatal Force. Retrieved from https://www.washingtonpost.com/graphics/investigations/police-shootings-database/
  • Giles, D., & Madsen, T. (2016). Big Data and Public Policy: Methods and Challenges. Journal of Policy Analytics.
  • Lynch, J. (2020). Data Visualization for Social Science. SAGE Publications.
  • R Core Team. (2023). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.
  • Miller, J., & Han, J. (2012). Input Data Quality and Its Role in Data Mining. Data Mining and Knowledge Discovery.
  • Kohavi, R., & Provost, F. (2002). Business data mining. Data Mining and Knowledge Discovery, 6(2), 107-118.
  • Cleveland, W. (1993). Visualizing Data. Hobart Press.
  • Schneider, T. (2018). Understanding and Using Data Visualization. Oxford University Press.
  • Yau, N. (2011). Data Points: Visualization That Means Something. Wiley.

Related Assignments