Housing Markets and Dimensionality Reduction
Professor: Erin Mayfield
Winter 2023
PROMPT
This analysis applies dimensionality reduction techniques, including principal component analysis and factor analysis, to discover actionable insight about the Hanover housing market using a dataset which observations for 691 housing units sold in the Upper Valley from 2019 to 2022, where each row of the dataset represents a housing unit.
Principal Component Analysis
Higher absolute values in the loadings matrix indicate a stronger relationship between the original variable and the corresponding principal component.
PC1: This component is positively associated with the number of bedrooms, number of bathrooms, square footage of indoor living space, and sale price. This suggests that larger and more expensive homes tend to be grouped together in this component. This component is negatively associated with the year the house was built, suggesting that older homes tend to be grouped together in this component as well.
PC2: This component is positively associated with the walkability rating of the neighborhood and negatively associated with the year the house was built. This suggests that newer homes in more walkable neighborhoods tend to be grouped together in this component.
PC3: This component is negatively associated with the average school rating in the area, suggesting that homes located in areas with lower-rated schools tend to be grouped together in this component.
PC4: This component is positively associated with the year the house was built and the walkability rating of the neighborhood. This suggests that newer homes in more walkable neighborhoods tend to be grouped together in this component.
PC5: This component is positively associated with the number of bedrooms and negatively associated with the sale price, suggesting that larger and more expensive homes with more bedrooms tend to be grouped together in this component.
PC6: This component is negatively associated with the number of bedrooms and the sale price, but positively associated with the number of bathrooms, suggesting that less expensive homes with fewer bedrooms but more bathrooms tend to be grouped together in this component.
