By examining air quality indices across the world we hoped to see some sort of relationship emerge that would allow us to monitor air quality and then use it to predict COVID infections. There was a very slight positive correlation between both general air quality and NOx in a country and COVID infection rates per million of population. However, the data on COVID worldwide is collected in different ways and on different populations of people, so our findings are unlikely to be a reliable predictor of COVID at this time.
My background (Karen McEwen) is in environmental engineering and data analysis, and my partner's (Gerald V. Alva) background is in computer programming. This is our first hacker challenge.
Since COVID-19 was first identified as a respiratory disease (though we now know it can present in many different ways), we wondered whether there might be correlation between countries with poor air quality and COVID-19 infection rates.
We started with space agency data from SEDAC, which detailed overall air quality in 180 countries, along with other environmental indicators, from 2018. We then found COVID data from 200+ countries in the world and took the most recent data (May 30, 2020). There were 171 countries in common between the two data sets. We used Excel to do initial cleaning of the data, python to merge the two data sets, and Tableau to plot the factors and look for trends.
Two air quality indicators showed a statistically significant positive correlation - Air Quality (overall) and NOx Emissions -and COVID cases per million. There was no statistically significant correlation between SO2 emissions and particulate matter (PM2.5) emissions.
Obviously, there will be many caveats, even where there is some correlation. And, of course, correlation does not imply causation. For example, not all countries are providing data at the same rate (daily) as other countries and not all places are in the same part of the infection cycle. COVID data is heavily reliant on testing, and where people are being tested at different rates, there will be a skewing of data when trying to compare countries. Also, the air quality data is an average over a year, from two years ago. It may be that we are comparing apples to bananas and oranges. However, it is a start.
In the future, we would have liked to do a dynamic look at air quality indicators and COVID cases. Our hope was to build an easy to use dynamic dashboard interface that would allow us to look at current data between air quality indicators and COVID cases, but that was beyond our time and resources.
NASA Socioeconomic Data and Applications Center (SEDAC)
Environmental Performance Index, 2018 Release
Citations:
Yale Center for Environmental Law and Policy - YCELP - Yale University, Yale Data-Driven Environmental Solutions Group - Yale University, Center for International Earth Science Information Network - CIESIN - Columbia University, and World Economic Forum - WEF. 2018. 2018 Environmental Performance Index (EPI). Palisades, NY: NASA Socioeconomic Data and Applications Center (SEDAC). https://doi.org/10.7927/H4X928CF. Accessed 30 May 2020.
Wendling, Z., D. Esty, J. Emerson, M. Levy, A. de Sherbinin, et al. 2018. The 2018 Environmental Performance Index Report. New Haven, CT: Yale Center for Environmental Law and Policy. https://epi.envirocenter.yale.edu/node/36476.
European Center for Disease Prevention and Control (ECDC)
https://ourworldindata.org/covid-cases
Citation:
Coronavirus (COVID-19) Cases Statistics and Research by Max Roser, Hannah Ritchie, Esteban Ortiz-Ospina and Joe Hasell - Accessed 30 May 2020.