Analysis of Socioeconomic Factors and Voting
Patterns in US Counties
Jeffery Delva
2025-09-24
1. Introduction
Objective & Background
The objective of this analysis is to explore the relationships
between 2020 presidential election outcomes, economic health
(unemployment and poverty), and education across U.S. counties.
By visualizing these variables and their interactions, we aim to
uncover patterns that can lead to more specific research questions.
The dataset used for this analysis combines county-level data from
four public sources: the MIT Election Lab, the Bureau of Labor
Statistics, and the U.S. Census Bureau.
2. Data Set Description
This section provides a summary of the dataset’s structure and
variables as required.
The dataset contains 3153 observations (counties)
and 11 variables.
Variable Definitions
The following table lists each variable and its description.
Variable Definitions for the Consolidated County
Dataset
| county_fips |
Numerical |
A unique 5-digit Federal Information Processing
Standard code for the county. |
| state |
Categorical |
The name of the U.S. state. |
| county_name |
Categorical |
The name of the county. |
| winning_party_votes |
Numerical |
Total votes for the winning party in the 2020
presidential election. |
| winning_party |
Categorical |
The political party that won the county in 2020. |
| unemployment_rate |
Numerical |
The county’s unemployment rate for 2020 (%). |
| PCTPOVALL_2019 |
Numerical |
The estimated percentage of people in poverty in the
county in 2019. |
| percent_less_than_hs |
Numerical |
The percentage of adults with less than a high school
diploma (2015-19). |
| percent_hs_only |
Numerical |
The percentage of adults with only a high school
diploma (2015-19). |
| percent_some_college |
Numerical |
The percentage of adults with some college or an
associate’s degree (2015-19). |
| percent_bachelors_or_higher |
Numerical |
The percentage of adults with a bachelor’s degree or
higher (2015-19). |
3. Exploratory Data Analysis
This section visualizes variable distributions and their
relationships to answer key analytical questions.
Analysis 1:
Distribution of Single Variables
Analytic Question: What are the typical
distributions of key economic indicators like poverty and unemployment
across U.S. counties?
Interpretation: The distribution of poverty rates is
right-skewed, meaning most counties have a poverty rate
below 20%, but a significant number of counties report higher rates,
indicating widespread economic disparity.
Analysis 2
Relationships Between Two
Variables
Analytic Question: Is there a relationship between a
county’s dominant political party and its level of educational
attainment?
Interpretation: The boxplot shows that counties won
by Democrats tend to have a higher median percentage of adults with a
bachelor’s degree or higher compared to counties won by Republicans. The
distribution for Democratic won counties is also more varied, with a
wider interquartile range and more outliers.
Analysis 3
Pairwise Comparisons
Analytic Question: How do all the key numerical
variables (education, poverty, unemployment) relate to one another?
Interpretation: The most distinct relationship is
the strong negative correlation between the percent_bachelors_or_higher
and the PCTPOVALL_2019 (poverty rate). As one goes up, the other clearly
goes down. There is a strong positive correlation between
percent_less_than_hs and the poverty rate, indicating that less educated
adults is associated with higher poverty.
4. Results & Discussion
The Exploratory Data Analysis reveals strong, consistent
relationships between education and economic status at the county level.
Counties with more highly educated populations tend to show lower rates
of poverty and unemployment.
Furthermore, these socioeconomic indicators appear to correlate with
political outcomes, as seen in the differing educational levels of
counties based on their 2020 election results.
Research questions that can be posed following this EDA consists
of: