Master of Science
Yunchuan Liu, Ph.D.
Xin (Jasmine) Chen, Ph.D.
Xueqing Tang, Ph.D.
This project presents a comprehensive analysis of COVID-19 data for Maharashtra, one of the most affected states in India. Utilizing advanced data processing techniques, the study focuses on understanding the progression of the pandemic through time series analysis and predictive modeling. Initially, the dataset, consisting of confirmed cases and deaths, is meticulously filtered to isolate data specific to Maharashtra. This subset is then transformed into NumPy arrays, facilitating the use of Python libraries for visualization and analysis.
The research employs a variety of visualization techniques, including line plots and histograms, to portray the temporal trends and frequency distributions of COVID-19 cases and fatalities. Key to the analysis is the implementation of a Long Short-Term Memory (LSTM) neural network, tailored to predict future trends based on historical data. The LSTM model is rigorously trained and tested, with its predictions plotted against actual data to assess its performance.
Furthermore, the study delves into the day-to-day changes in the pandemic data by calculating first differences, providing insights into the daily fluctuations in cases and deaths. These differences are visually represented through scatter plots and histograms, elucidating the pattern of peaks and troughs indicative of the pandemic's dynamics.
The outcomes of the analysis are twofold: a detailed understanding of the COVID-19 trend in Maharashtra and a predictive model capable of forecasting future case and death counts. The model's predictions, alongside the visualizations, offer critical insights into the progression and potential future trajectory of the pandemic. The findings of this study are significant for policymakers and healthcare authorities, aiding in strategizing effective responses to mitigate the impact of the pandemic in Maharashtra and potentially in similar regions.
Kusanapalli, Manikanta, "COVID-19 Trends in India: Nationwide Insights & Maharashtra Focus" (2023). All Capstone Projects. 687.