Flu Trends vs Outpatient Visits

This project explores the correlation between Google Trends searches for "flu symptoms" and CDC Influenza-Like Illness (ILI) outpatient visits, to see if an accurate linear regression model can be used to predict ILI visits from Google searches.

---

Project Overview

• Goal: Examine whether Google Trends data can predict flu activity and how COVID-19 changed that relationship.
• Data sources:
  • Google Trends: Searches for "flu symptoms" in the US from 2014–2026 using the gtrendsR package
  • CDC ILI Data: Weekly outpatient visit data aggregated by month (2015–2026)
• Analysis:
  • Linear regression on aggregated data
  • Train/test split to evaluate predictive performance
  • Correlation comparisons pre- and post-COVID
• Visualization: Trends over time of ILI visits and Google trends, scatterplot comparing search hits and mean ILI activity, plotted ML model with LOESS smoother.

---

How to View the Report

The full report contains all code, plots, and analysis. It is rendered in HTML using Quarto and hosted via GitHub Pages:

View the live report

---

File Structure

• scripts/ — R scripts and Quarto .qmd file
• index.html — rendered report (GitHub Pages)

---

Notes

• The HTML report is self-contained, so it can be opened locally without missing assets.
• All analysis was performed in R using tidyverse, ggplot2, lubridate, and gtrendsR.