Are DIA Airline Passenger Counts Affected by CO COVID-19 Case Counts from 2020?
The COVID-19 pandemic has had many short-term and long lasting effects on how we live our daily lives, specifically how we travel. One of the sectors of transportation that took a hit due to coronavirus was the air transportation industry. The frequency of flights during the 2020 year of the pandemic has been studied and has proven to be quite a costly disruption. In addition to the frequency of flights, the quantity of passengers is another factor of traveling that could have possibly been affected by the pandemic. With good reason, one could hypothesize that flyers may not have been able to fly due to the travel bans that were imposed and most were not as apt to hop on a plane in light of the spread of the novel coronavirus. These concerns of safety and well-being of air passengers in the midst of COVID-19 is an interesting subject that may have influenced the amount of air travelers in 2020.
This analysis attempts to explore if there is any relationship between the number of flight passengers from March-December 2020 for the Denver International Airport (DIA) and the total number of COVID-19 cases for Colorado. The number of coronavirus cases were aggregated from county level data in the state of Colorado. The airlines that were considered in this analysis were: American, Delta, Frontier, Other, Southwest, United, and United Express.
The distribution of flight passengers for each airline appeared to be skewed left. This may suggest that most of the data observations (flight passenger count) were more frequently of higher counts. This distribution holds true for the majority of the airlines, with some variations of some data leaning closer to zero, but still having a large uptick in the frequency of passenger count toward the right side of the horizontal axis. The COVID-19 cases data on the state level of Colorado appeared to be more right skewed, meaning that most data observations were closer to zero on the horizontal axis. Therefore, there was a higher frequency of data that were instances of smaller amounts of statewide total monthly cases. Just from these visual estimations, one could say that most data observations for the airline data were far from zero and the observations from the COVID-19 cases were closer to zero. Meaning, there were generally more instances of smaller amounts of cases and more instances of larger passenger counts.
Pearson’s correlation coefficient tests were conducted for each of the airlines and the COVID-19 cases data. The correlation coefficient test results for the airlines, American, Delta, Frontier, Other, Southwest, United, and United Express, were 0.27, 0.22, 0.26, 0.11, 0.27, 0.48 and 0.41, respectively. The airline that corresponded to the highest correlation coefficient, 0.48, was United Airlines. This relationship did not visually appear to be very correlated with the COVID-19 cases data in a line plot, and the correlation coefficient of 0.48 reflects this observation. The p-values for the airlines respectively are: 0.45, 0.55, 0.46, 0.77, 0.45, 0.16, and 0.24. The issue with performing a test statistic with this particular dataset is that there are not many samples, which can significantly reduce the reliability of the p-value. When there are less samples, it may not fully present the picture of the phenomenon at hand, and therefore may provide a false correlation result.
Given the sources of uncertainty that were just identified, what can be said about the correlation (if any) between the variables described? At least, we can say that with the tests that were conducted, the relationships appeared to be positively correlated. It would be misleading to make any firm determination about the statistical relationship between monthly flight passengers and the number of monthly COVID-19 case numbers for Colorado, due to the high uncertainty present.
Oftentimes, when there are interesting questions to be asked about real world phenomenons, the results may not confirm our hypotheses and there will more than likely be instances of uncertainty. However, this analysis did show that there is more exploratory studies to be done to see if these variables are related. Furthermore, the addition of other explanatory variables may shed some light on this story, and possibly show a relationship between flight passenger count and COVID-19 cases.
