Situation:
COVID-19 caused millions of deaths worldwide, and one of the biggest challenges was that many infected individuals showed no symptoms. Early detection methods could help identify outbreaks faster and guide public health responses. Our team wanted to explore innovative, data-driven approaches to detect potential COVID-19 spikes in different regions of the U.S.

Task:
Develop a system that integrates wastewater viral load data, Google Trends search activity, and machine learning models to detect early signs of COVID-19 spread. The aim was to build a proof-of-concept dashboard for real-time monitoring and forecasting.

Action:

• Data Collection: Gathered wastewater viral load datasets from U.S. counties and integrated Google Trends data via the Pytrends API to capture search behavior patterns (e.g., “loss of taste,” “COVID testing”).
• Exploratory Analysis: Used Python (Pandas, Pytrend) and R for data cleaning, preprocessing, and correlation studies.
• Machine Learning:
  • Implemented Deep Neural Networks to model relationships between viral loads, search activity, and case counts.
  • Designed experiments in Jupyter Notebook and Google Colab for scalability.
• Application Development: Prototyped front-end options (Flask, Dash, RShiny, PowerBI) and tested backend integration for a dashboard to display predicted COVID-19 spikes.

Result:

• Built a proof-of-concept early warning system combining environmental data and digital search behavior.
• Demonstrated feasibility of using Google Trends + wastewater data as a leading indicator for outbreak detection.
• Provided a scalable framework that, with refinement, could support public health agencies in proactive COVID-19 response and future pandemic preparedness.

The above image is an example of predictive analysis to predict the chances of COVID-19 based on the information provided