Impact of Covid-19 on Mobility, Air Traffic and Automobile Accidents

Team165, CSE 6242: Data and Visual Analytics

Following dashbord is the interactive visualization prepared on tableau.
The dashboard explores the relationsip between covid-19 cases and human mobility, air traffic flow and automobile accidents.
Read further below at the about section on our motivation, data, sources and experiments.
The dashboard is also directly available at Tableau public.

Instruction to Use the Dashboard

  1. Situated at the center of the dashbord is the map of US
  2. Map Navigation
    • Hover over the map to see details of each state
    • Click and hover on a state to see the state's county level view
    • To reset the map or get out of current view simply click 'Go Back to US View' on top left corner
  3. Comprehensive Information Box
    • Located right on top of US map
    • Provides comprehensive total of covid cases, accidents, flights and movility index for the US by default
    • View changes to state-wise data when clicked on a state.
  4. Metric Trend Box
    • Located to the left of US view
    • Provides trend-lines of metrics for US by default
    • Trend-lines changes to state when clicked on a state
  5. Experiments
    • Located at the bottom of Dashboard
    • This view shows the Regression Analysis done with 2 weeks lag period
    • The metrics to determine relationship and statistical significance were Beta, Correlation, R-Squared, T-Test and P-values
  6. Navigator and Legend
    • Located to the right of the map
    • Click on the date bar to select date to view the map for.
    • To automatically play/stop map from start date to end date, click on the play/stop icon.

About the Project

Its been alomst more than 1 and a half years the novel covid-19 virus discovered. Its outbreak has made some serious impacts on all aspects of life, countless infections and numerous deaths.
To mititgate the risk multiple cautions and restricitions have been placed. We the members of team165 wanted to see if there was any relationsip between human mobility and the rise and fall of covid-19 cases. Our solution was to build a visualization for the relationsip Analysis The Analysis we belive would tell us if increase or decrease in covid cases had any effect on flights, mobility and automobile accidents. For our experiment we did extensive research for reliable data. We used John Hopkins University's covid-19 database to obtain data for covid-19.
For automobile accidents we used freely available dataset from Kaggle which was maintained by data scientiests from Lyft. FAA's airport operation report generator and ndfc facility generator came handy to find out daily domestic flight in each of US airport. Finally, we used Google's publicly available mobility dataset. Total size of the data was around 5 GB. The data required thorough cleaning.
After cleaning data and reducing the size, we had around 1 million rows of data to work with. Our algorithm followed google's mobility index. Where each weekday would be baselined using median from first 5 week of the year. The percentage change from baseline would be the weekly socre. Weekly socre represents the relationship between metric of calculations such as increase in mobility increased in covid-19 cases. Regression alayisi was also done for experiment. A common hypothesis was increase in covid-19 deaths would be associated with lower accident levels.
We regressed accidents on COVID-19 deaths. The results did not support our hypothesis. Furher experient analysis can be viewed on the dashboard

Meet the Team!

Avatar

Brett Durborow

Engineer

Avatar

Jeewan R. Bhattrai

Engineer

Avatar

Kriti Acharya

Engineer

Avatar

Mackenzie Schaich

Engineer

Avatar

Michael Spagnolo

Engineer

Avatar

Ribash Sharma

Engineer

Contact Us with Suggestions!

Your Message Has Been Sent!