Page 67 - FULL REPORT 30012024

2.9.5 Data Visualization and Analysis with Machine Learning for

the USA’s COVID-19 Prediction

The proposed research presents a web-based COVID-19 information hub that
aims to provide users with intuitive data analytics and visualization features,

as well as a chatbot for answering frequently asked questions. The system
utilizes machine learning techniques such as Linear Regression (LR) and

Support Vector Machine (SVM) for COVID-19 data analysis and prediction.

Data from multiple sources, including the COVID-19 Github Dataset and the
World Health Organization (WHO) Question and Answer page, are collected

and pre-processed for analysis.

The system's data visualization capabilities include displaying and

visualizing daily and cumulative COVID-19 cases, deaths, and vaccination
statuses at both global and state levels. It also offers pandemic trend

predictions for cases, deaths, and vaccinations, allowing users to explore the
past 30 days of data and the next 30 days' predictions. Additionally, the

system provides the latest news articles related to COVID-19 in the United
States, along with a chatbot feature trained on WHO data to answer user

questions.

The overall system design includes components such as a Python system for

data collection, analysis, and machine learning, the Dialogflow platform for
NLP and chatbot training, a Node.js RESTful API server for communication

between the frontend and backend, and an Angular-based web system for user

interaction. MongoDB is used as the database for storing complex case data
and prediction results.

Experimental results demonstrate that SVM outperforms LR in terms of

prediction accuracy, as indicated by the mean squared error calculation.
Therefore, the system selects SVM as the preferred machine learning

algorithm for the final product.

50