Introduction

After being sent home from my sophomore year of university in March of 2020 due to the COVID-19 pandemic, I joined my family, friends, and neighbors for walks more often than ever before. It was a great way to stay connected in person, while still being able to wear a mask and distance ourselves 6 feet apart. The restrictions in my hometown of Virginia Beach, VA were limited, so going out for a walk was possible. 

It seemed as though everyone was going out walking to take a break from the Zoom calls, TikToks, banana bread, and Dalgona whipped coffee. However I wasn’t in contact with nearly as many people as prior to March of 2020, so my scope was just the people I was going out for walks with or the occasional FaceTimes from other active friends.

So I sought a way to find out what the activity during the pandemic looked like globally. Finding the Apple Mobility Trends data set intrigued me, it would showcase not only walking data but driving and transit as well. This would be a good comparison, and made me ask the following questions: 

1. Could walking potentially overtake driving in comparison to pre-pandemic? 
2. When breaking it down the country by country- did all countries see an increase in walking or mobility? 
3. How much had the use of transit decreased? And has it returned to the baseline?
4. What events could have affected the mobility scores? 

When thinking back to what the main activity controller was from 2020-2021 it typically revolved around case numbers. Lower numbers meant more to go do whereas higher numbers meant shutdowns and cancelations. So to compare dips in mobility I looked to find a data set of new cases, hypothesizing that higher cases would mean lower mobility, especially since the mobility data ended in October of 2021, before the rise in Omicron Variant cases, which were less severe and with vaccination had shorter quarantine times. The data set I pulled was from “Our World in Data”, which was able to provide me with new case numbers daily, I selected to have it broken down by continent, only to later decide to look at new cases as a whole.  

With these two data sets, I would be looking to see if there was an inverse relationship between the amount of Covid cases and the mobility score around the world. Meaning that if the number of new Covid cases was up, then mobility would fall below the baseline and if new cases were down, mobility would be above baseline. 

Diving deeper into the Apple Mobility Trend Data Set I would like to pull the countries that had the greatest increase in walking during the time frame and the countries that experienced the greatest decrease in walking over the course of the 22-month period. This could lead to other ideas of how these mobility scores were affected. I suspect that the strictness of lockdown could potentially have an effect on how the countries rank. During the pandemic, some countries put strict curfews or didn’t allow people to leave their residential areas, whereas others proceeded with life as normal. 

Inspiration

When looking into data sets that contained data from 2020-2022 on people’s activity I came across this article titled COVID-19 creates new momentum for cycling and walking. We can’t let it go to waste! In this article, they included the chart above. I was very intrigued to see how the data looked now that it had been 2 years and see where the walking, driving, and transit trend lines were moving.

Process

The data set I first utilized was Apple’s Mobility Trend report, which can be found here linked below on Kaggle. Apple has since discontinued these trend reports as of April of 2022. The data set here consists of data from January 2020 to October 2021.  

Apple Mobility Trend Report: January 2020-October 2021

Using Open Refine I transposed the data so that the mobility data could be observed over time. After transposing the data the set contains 3,025,971 data points containing the daily mobility scores for driving, walking, and transit from Apple Maps from 63 countries.

The next data set I chose for COVID-19 new cases and total cases came from Our World in Data. I pulled this file by continent to see trends in that area. I filtered the data set down to the same date range as the Apple Mobility Trends. 

Our World in Data: Covid Case Data

 From here I loaded both data sets into Tableau. My original plan was to create maps, with pages of each month globally, then by continent, and color the countries based on their mobility score. These were turned into GIFs. I used only the 2020 data so that the GIFs would have 12 months and then repeat. I used these in my first UX participant test asking them to interpret and talk about any trends they saw while viewing the GIF.

GIF Ideas (May take a few moments to load):

User Testing Results

This proved to be an unsuccessful way of showcasing mobility as the user said it moved too quickly, wasn’t pausable, and the user couldn’t understand the graph. 

Going into my next user test I decided to present it in a poster format, focusing on displaying each map by monthly average and by seasonal average. With this poster, I included the number of Covid cases in a bubble chart format as well. (Disclaimer the poster below is not my final, but what was shown to the user during testing)

I asked the user to find:

1. Average Covid Case Numbers in North America
2. The baseline mobility score
3. What season transit had the least amount of use
4. What month had the lowest mobility overall

The user was able to identify each but suggested that the data be available to drill down into so that we could see each continent individually. They also recommended a line graph to display each of the transportation types over time. I agreed that that was a great idea, especially since the map version really hadn’t been translated in the first user test. 

The best feedback I got though was in class. Neither one of my participants had mentioned the colors- and I had gone a little overboard with the colors. The continent’s color-coding conflicted with the mobility with yellow, orange, and green being used for both. The entire poster was overwhelming. 

Taking this into account I went back to the drawing board. I wanted to create something that told the story of the rises and falls in mobility. I ditched the colorful maps. They were fun, but not understandable. 

Final Poster

PDF Version:

I decided to make a simplistic poster, that would incorporate maps to indicate location, but layer it with a line graph to showcase the mobility. The line graph at the top features each mode of transportation from the data set: driving, walking, and transit. In order to draw attention to what was happening in the world at points in the graph, I googled “news” and set the date range to that month. Then I highlighted the following news articles at those points in time: 

April 2020: https://www.bbc.com/news/world-52103747

August 2020: https://bjsm.bmj.com/content/54/20/1183

January 2021: https://inews.co.uk/news/environment/cycling-walking-much-more-popular-covid-19-lockdown-849025

July 2021: https://www.theguardian.com/us-news/2021/jul/18/coronavirus-delta-anxiety-reopening

I also included discoveries from the data set directly, like when transit returns to pre-pandemic mobility levels or when walking passes driving in having the highest mobility scores.

The bubble graph below displays the Covid cases along the same timeline as the line graph in order to highlight any sort of correlation between the two. I called out one range directly on the chart where we see a rise in new Covid cases and a dip in mobility. 

After mapping the data earlier I had also learned that Apple did not have data from every country in the world, so I made sure to list out which countries were included in this data set off to the right-hand side. 

The final set of graphs includes four bar graphs representing the walking scores for four specifically selected countries. These countries are the ones with the highest increase in walking mobility in comparison to the baseline and the ones with the highest decrease in walking mobility in comparison to the baseline. I used the same scale of colors for each of these graphs, going from a mobility score of 20 being red, centering the yellow coloration at 100, and making the top value green. (Anything 200+ was the same color of green since Croatia’s graph ranged to upwards of 700.) This showcased the large difference between countries, especially since in the global line graph walking stays above baseline for the majority of the time, and it indicates how different the overall pandemic experience was across the globe. 

Results

The trend of mobility did not appear to always follow new covid cases. At the beginning of the Pandemic, the new case numbers were lowest and so was mobility overall. This could be due to the high number of lockdowns during this time. One of the limitations of Apple’s data is that it doesn’t show us physical activity, only if a person looked how to walk to a certain place. It’s possible that physical activity or working out at home may have increased during this time, however, this data is unable to show that. 

From October 2020-March 2021 we do see a dip in overall mobility as new case numbers are at their highest for the time range of January 2020-October 2021. The hypothesis that the mobility score would drop below baseline is not able to be seen as although driving and walking fall, they are still above the pre-pandemic baseline and nowhere near the low numbers seen in April of 2020. Transit remains under baseline up until June of 2021, where for the first time since March of 2020 it returns to pre-pandemic levels or higher. 

One surprising finding was that walking did not pass driving in mobility score until August of 2021. I had expected walking to increase the most from the baseline originally and then turn down once people began to return to in-person work or school. However, the trend line showed that driving was consistently above the baseline from May of 2020 onward. Thinking about it now, perhaps this could be due to road trips or other activities involving needing directions that many wouldn’t have been able to do prior to working remote. Another limitation with the data set comes in here- although people may be driving more, they might not be looking up directions on Apple Maps to get to work, therefore lowering the mobility score. 

For the walking assessment, two countries stood out with top mobility scores. Croatia and United Arab Emirates (UAE) had the highest by far. Croatia is by far the highest outlier on the chart. Looking into Croatia’s government’s actions during the pandemic, they stayed focused on continuing tourism and their economy. They opened borders to Americans, going against the European Union’s decision to restrict U.S. travelers. Being that they were one of the few countries open to visitors, makes sense why the huge rise in walking mobility scores only during the summer months. Other countries that scored above 250 (in comparison to the 100 baselines) include Albania, Slovenia, Bulgaria, Slovenia, Estonia, Spain, and Lithuania. All of these countries, besides UAE, are in Europe. 

After removing Croatia and UAE from the line chart I was able to see that Macau and Thailand trailed below the rest when it came to mobility scores. Throughout the time frame, Macau had never come back to the baseline. (Thailand was only above baseline in January of 2020.) Out of the 63 countries apart of this data set, 9 had mobility scores under the baseline of 100 consistently from February 2020-October 2021. This includes Cambodia, Hong Kong, Indonesia, Malaysia, New Zealand, Philippines, Republic of Korea, Singapore, and Uruguay. All of these countries, besides Uruguay & New Zealand, are in Asia. 

Taking note of the two continents, we can refer back to the COVID case data to see the ranking of the top number of COVID Cases during this time frame. 

1. Asia (~16 Billion)
2. Europe (~15 Billion)
3. North America (~14 Billion)
4. South America (~10 Billion)
5. Africa (~2 Billion)
6. Oceania (~38 Million)

The continent of Asia had the highest total number of cases, perhaps contributing to more lockdowns, and less walking potential for residents, in comparison to Europe, where despite having the 2nd highest number of total cases lockdowns were less common and tourism was welcomed back relatively quickly. 

After comparing this I would conclude that the strictness of lockdown had the greatest effect on the walking mobility scores of a country. I would like to compare these data points in the future using the University of Oxford’s data set that tracks the government’s response to Covid-19. 

https://www.bsg.ox.ac.uk/research/research-projects/covid-19-government-response-tracker

The time spent during the pandemic is a tricky one to figure out. There are so many factors that go into results. What correlates to the world’s mobility is a compilation of lockdowns, restrictions, weather, lifestyle, and much more. The causation and correlation will be something I continue to dive deeper into with this. By adding in more factors perhaps something will lead the way in describing what caused the responses in mobility.