Visualization and Machine Learning Prediction of Spatiotemporal Spread of COVID-19 in India

Abstract

The global COVID-19 pandemic, affecting over 8 million people across 100 nations, presents a severe risk to human life and property. India, with its vast population of 1.34 billion, is among the hardest-hit countries. This study employs machine learning techniques for visualizing and predicting the spatiotemporal progression of COVID-19. Utilizing Python libraries such as “Numpy,” “Pandas,” “Scikit,” and “Matplotlib,” we analyze and visualize COVID-19 data sourced from the Indian Ministry of Health Web Service and API. Our visualizations depict demographic trends, incident growth, geospatial state-based patterns, and distribution. The analysis reveals that age groups under 30 and over 59 exhibit resilience to the virus, offering hope for population growth. Examining active cases, recoveries, and deaths, India has outpaced countries like Germany, the United States, Iran, Italy, Spain, South Korea, Turkey, France, and the United Kingdom since early April 2020. Furthermore, we employ supervised machine learning algorithms, including PROPHET and ARIMA, to predict the virus’s spread. By accounting for seasonality-related factors, we achieve a 95% prediction interval, indicating the potential for accurate spread forecasting. This research contributes valuable insights into COVID-19’s impact in India and offers predictive tools for managing its progression.