WITHRAWN: Identifying County Demographics Most Affected by COVID-19

• Activity Number: 75 - SPEED: Data Challenge and SPAAC
• Type: Topic-Contributed
• Date/Time: Monday, August 9, 2021 : 10:00 AM to 11:50 AM
• Sponsor: Section on Statistical Computing
• Author(s): Eric Christopher Browne and Isabel Osgood
• Companies: University of Denver and University of Denver
• Keywords: Linear Regression; Clustering; Agglomerative Nesting; County; COVID; Ward's Method

Abstract

The theme for the 2021 ASA Data Challenge Exposition is that of “Helping Families, Businesses, and Communities Respond to COVID-19”. We believe that the best way to help the citizens of the United States respond to the pandemic is to properly invest in its demographics that are most vulnerable to the pandemic. Through clustering techniques such as Agglomerative Nesting and Divisive Hierarchical, we identified underlying structures in U.S. counties, sorting them based on our response variable: Covid Impact. Covid Impact is defined as the sum of the COVID-19 cases and deaths in a county; standardized on population. The success of these clusters was measured using Linear Regression and a Tree Regressor in identifying key factors associated with high covid impacts. The most important factors in predicting Covid Impact for counties through Linear Regression besides cluster labels were the percentage of population that identifies as a minority, the percentile of unemployed, and the percentage of occupied housing units with more people than rooms. Our proposal to the United States legislature is to properly invest in and prioritize relief aid for these demographics.