COVID-19 Case and Excess Mortality Analysis
Published:
Analysis of the COVID-19 Case Rate and Excess Mortality Rate by Pandemic Wave
๐ COVID-19 Pandemic Wave Analysis in the U.S.
Team Members: Shuoyuan Gao, Yiqiao Zhu
Course: BIOS 620 โ Introduction to Health Data Science
Institution: University of Michigan School of Public Health
GitHub Repository: Analysis-of-the-COVID-19
๐ Project Overview
This project analyzes the evolution of COVID-19 case and excess mortality rates across three major pandemic waves in the United States from January 2020 to April 2025. Using state-level weekly data on confirmed cases, total deaths, and expected deaths, we built models to quantify and forecast the impact of the pandemic.
We investigated:
- Wave identification using time series plots of case, death, and excess mortality rates.
- State-level variation in severity during each wave.
- Modeling virulence using case fatality rates (CFR).
- Cross-wave prediction using linear, LOESS, and spline models.
๐ Key Findings
- Wave 1 (MarโJun 2020): Highest CFR (4.5%) and Northeast hardest hit.
- Wave 2 (Oct 2020โFeb 2021): Deadliest wave with widespread impact.
- Wave 3 (JulโOct 2021): Lower CFR due to improved healthcare and vaccination, but regional mortality remained high.
Model Insight: Nonlinear models (e.g., splines, LOESS) outperformed linear models in predicting excess mortality, especially when trained on one wave and tested on another.
๐ Report Access
๐ Final Report (PDF) ๐๏ธ GitHub Repository
๐ Wave-specific metrics, mortality maps, CFR charts, and model comparison plots included in the full report.
๐งช Tools & Methods
- Languages & Libraries: R,
ggplot2,dplyr,splines,loess - Data Sources:
๐ง Acknowledgements
This project was completed for BIOS 620 at the University of Michigan. We thank Professor Dylan Cable and our GSI Yize Hao for their support and guidance.