Exploring the Spatial Relationship Between Housing Cost and Family Households in New York City - Packet Flow: Journey to Network & Cybersecurity Expertise

Abstract

This study examines the relationship between median rent in 2008 and the percentage of households with children in 2008 across New York City sub-boroughs. Exploratory spatial data analysis was conducted using GeoDa to assess the distribution of data, identify outliers, and evaluate statistical relationships among variables. Results indicate that median rent is positively skewed, while the percentage of households with children is more evenly distributed. A statistically significant negative association was identified between rent and the presence of family members in the household. A bivariate choropleth map created in ArcGIS Pro visually confirms that higher-rent neighborhoods tend to have lower proportions of households with children. The integration of statistical analysis and spatial visualization enhances the interpretation of socioeconomic patterns in urban environments.

Introduction

This project investigates whether housing cost is associated with household composition in New York City. Specifically, it examines the relationship between median rent (2008) and the percentage of households with children (2008). The analysis integrates exploratory spatial data analysis in GeoDa with cartographic visualization in ArcGIS Pro.

Understanding distributional patterns and statistical relationships prior to mapping is essential in spatial analysis. This approach ensures that geographic interpretations are grounded in quantitative evidence.

GeoDa and the Exploratory Spatial Data Analysis Framework

GeoDa, developed by the Spatial Analysis Laboratory at the University of Chicago, is an open-source software designed to support exploratory spatial data analysis. Its purpose is to help analysts understand spatial datasets through interactive statistical graphics linked to maps.

In this project, GeoDa was used to generate histograms, box plots, and scatter plots for the variables rent2008 and kids2008. The software’s linked selection feature allowed statistical observations to be connected directly to geographic locations. This step provided a clear understanding of distributional characteristics and relationships before producing final maps in ArcGIS Pro.

Exploratory Analysis in GeoDa

Distribution of Median Rent (2008)

The histogram of rent2008 shows that most sub-boroughs have rents between approximately 800 and 1,600 dollars, with a smaller number of neighborhoods reaching nearly 2,900 dollars.

*Figure 1. Histogram of Median Rent (2008) showing right-skewed distribution.*

The box plot confirms that rent is positively skewed. The mean rent ($1,257) exceeds the median ($1,100), indicating that high-rent outliers influence the distribution.

*Figure 2. Box Plot of Median Rent (2008) highlighting upper outliers.*

Distribution of Households with Children (2008)

The histogram of kids2008 shows values ranging from approximately 8 percent to nearly 50 percent, with most neighborhoods falling between 25 and 45 percent.

*Figure 3. Histogram of Percentage of Households with Children (2008).*

The box plot indicates a relatively balanced distribution. The mean and median values are close, suggesting limited skewness.

*Figure 4. Box Plot of Percentage of Households with Children (2008).*

Statistical Relationship

A scatter plot was used to examine the relationship between rent2008 and kids2008. The regression line slopes downward, indicating a negative relationship. The R² value of approximately 0.55 suggests a moderately strong association.

*Figure 5. Scatter Plot of Median Rent vs. Percentage of Households with Children (2008).*

The statistical results indicate that as rent increases, the percentage of households with children decreases.

Bivariate Choropleth Mapping in ArcGIS Pro

Following the GeoDa analysis, a bivariate choropleth map was created in ArcGIS Pro 3.6.1. Both variables were classified into three categories using a 3 × 3 grid. This resulted in nine combined classes representing the interaction between rent and household composition.

*Figure 6. Bivariate Choropleth Map of Median Rent (2008) and Percentage of Households with Children (2008), NYC Sub-Boroughs.*

The map reveals a clear spatial pattern. High-rent areas are generally associated with lower percentages of households with children, particularly in central Manhattan. Lower-rent areas tend to correspond with higher proportions of family households, especially in outer borough neighborhoods.

An inset map of New York State was included to provide geographic context.

Discussion

The integration of GeoDa exploratory analysis with ArcGIS Pro mapping demonstrates the value of combining statistical and spatial methods. The skewed distribution of rent and the balanced distribution of households with children suggest different underlying socio-economic dynamics. The statistically significant negative relationship indicates that housing costs are associated with household distribution.

Mapping both variables simultaneously in a bivariate choropleth strengthens interpretation by anchoring statistical results in geographic space.

Conclusion

This study illustrates how exploratory spatial data analysis and cartographic visualization can be integrated to reveal meaningful urban patterns. GeoDa provided insight into distribution, outliers, and statistical relationships. ArcGIS Pro translated these findings into a spatial representation that clearly shows the interaction between housing cost and family households.

Together, the analysis demonstrates that higher-rent areas in New York City tend to have lower proportions of households with children. Integrating statistical analysis with spatial mapping produces a clearer and more defensible interpretation of socio-economic patterns.