Predicting mobility using limited data during early stages of a pandemic

Area of Science:

• Behavioral Economics
• Epidemiology
• Data Science

Background:

• The COVID-19 pandemic profoundly altered consumer behavior and mobility patterns globally.
• Understanding these shifts is crucial for urban planning and public health strategies.
• Perceived risk associated with the pandemic significantly influences individual movement decisions.

Purpose of the Study:

• To investigate the key drivers of consumer mobility in US metropolitan areas during the COVID-19 pandemic.
• To analyze the impact of perceived risks, including disease prevalence and personal protective equipment (PPE) trends, on mobility.
• To develop and apply methods for mitigating multicollinearity in models using Google Trends data.

Main Methods:

• Utilized local and national COVID-19 case and death counts to quantify perceived risk.
• Incorporated real-time Google Trends data for personal protective equipment (PPE) searches.
• Developed a novel optimization scheme to address multicollinearity and identify significant covariates in linear models.

Main Results:

• Identified distinct factors driving mobility in different metropolitan areas.
• Discovered several common factors influencing mobility across all studied locations.
• Demonstrated the effectiveness of the novel optimization scheme in handling multicollinearity from Google Trends data.

Conclusions:

• Consumer mobility during the pandemic is influenced by a complex interplay of location-specific and general factors.
• The developed methodology provides a robust approach for analyzing mobility data influenced by public health crises.
• Findings offer insights for policymakers and businesses adapting to pandemic-induced behavioral changes.