N

Neighboring State Analysis for Covid-19 Cases

Computer Science Department

Kent State University

Project Description

Coronavirus has claimed millions of lives across the globe, since its widespread emergence in 2020. CDC COVID-19 Spatio-Temporal dataset for state cases was used to find correlations between states overtime. Can the analysis of COVID-19 trends and relations between neighboring states help predict the spread and prevent the loss of lives? In this project, we measured the correlation of the trends between neighboring states and created dynamic visualizations to show the strengths of the connections between states over time. With the help of these relationships, the spread of COVID-19 can be analyzed and predicted from one state to another.

Plotly Dash interface is used to visualize correlation of COVID-19 cases compared between neighboring states in order to find trends to help predict the spread of COVID-19. Weekly, Monthly, Bi-Annually, and total windows are analyzed. Cytoscape is used to create a network graph over time of nodes (states) and edges (correlation). This project was for Choose Ohio First and was presented on April 10^th, 2022.

Contributions

o Used Python to identify trends of cases per state

o A graph was created to visualize the states and connections to their neighbors.

o States visualized spatially with the percent change in cases displayed as a color

o Correlation value displayed as a colored node to display a positive or negative change.

o Correlation strength displayed as edge thickness

o Animated temporal graph created for different time frames

Documentation

Data Sets: CDC (https://data.cdc.gov/Case-Surveillance/United-States-COVID-19-Cases-and-Deaths-by-State-o/9mfq-cb36/data)

Source Code: GitHub (https://github.com/brandoncossin/COVID-19-Data-Mining)

Poster: http://www.cs.kent.edu/~xlian/projects/COF2022_Covid_Vis/CossinTothCOF.pdf

Video Presentation: https://www.youtube.com/watch?v=TF-fyagA3qE

Visualization Demo

States	Correlations in June 2020	Correlations in June 2021
Indiana	.421	.984
Michigan	.746	.966
Pennsylvania	-.314	.993
West Virginia	-.787	.965
Kentucky	.414	.989

Table 1: Case Study: Pearson's Correlation Coefficients Between Neighboring States and Ohio

Group Members

Brandon Cossin

Computer Science major class of 2022 with interest in data science and full stack development.

Email: bcossin@kent.edu

Troy Toth

Email: ttoth5@kent.edu

Dr. Xiang Lian (Advisor)

Email: xlian@kent.edu

Homepage: http://www.cs.kent.edu/~xlian/index.html

Last Modified: 4/28/2022