COVID-19 in India: Statewise Analysis and Prediction

Area of Science:

• Epidemiology
• Public Health
• Data Science

Background:

• The COVID-19 pandemic, originating in Wuhan, China, rapidly spread globally, reaching India in January 2020.
• India reported over 37,000 COVID-19 cases by May 3, 2020, with a rapidly increasing trend.

Purpose of the Study:

• To analyze state-wise COVID-19 infection data in India.
• To predict future infection numbers for each state over the next 30 days.
• To aid state governments in optimizing healthcare resource allocation.

Main Methods:

• Utilized logistic, exponential, and susceptible-infectious-susceptible models for prediction.
• Developed an ensemble model combining logistic and exponential predictions.
• Employed the maximum daily infection rate (DIR) over two weeks as a weighting factor and trend indicator.

Main Results:

• Categorized Indian states into severe (e.g., Maharashtra, Delhi, Gujarat), moderate (e.g., Tamil Nadu, Rajasthan), and controlled (e.g., Kerala, Haryana) infection levels.
• Achieved R-squared values above 0.90 for logistic and exponential models, indicating good fit.
• Provided a web application for regularly updated COVID-19 forecasts.

Conclusions:

• States with non-decreasing daily infection rates require intensified preventive measures.
• States with decreasing daily infection rates can maintain current measures, aiming for sustained negative rates to declare the pandemic's end.