Apple Maps — Mapping Mobility During COVID-19 Crisis
Last week, in the midst of a group project for my data science class, I took a few minutes to watch Morning Joe on MSNBC. These days, between election news and COVID-19 news, it seems like data science is talked about every day. My project was about the sentiment of citizens on COVID-19, so my ears pricked up when I heard Donny Deutsch refer to a COVID-19 mobility study from Apple.
I immediately went to the web where I found the Apple Maps Mobility Trends Reports, complete with a downloadable CSV file. While we were only able to use a fraction of the data in this file, there is a lot of interesting information here. The basic concept was this:
January 13, 2020 was set as a baseline, with travel for that day being represented by 100 (you could argue that they could have used a weekly average as the baseline but beggars can’t be choosers)
Each day after that is represented by a number which is relative for the usage of January 13, 2020
o 105 would be 5% higher
o 95 would be 5% lower
This method was applied to countries, regions (including states) and sub-regions. For several of these, data was included for different types of travel:
o Driving
o Transit (public)
o Walking
After importing the Apple file into python, I was able to plot US data for all three types of travel together on a graph. Not surprisingly, it showed that mass transit mobility has decreased the most — it is at around 20% from where it started and it has been pretty flat since late March. The numbers for walking and driving mobility dipped sharply in mid-March but they have been steadily rising since the beginning of April.
The study I was working on for school was looking at COVID-19 data for New York and Texas. I was only able to use the statewide data for the study but I was really fascinated by the data that I found for five different cities in Texas.
The map below shows the driving data for each city and we observe a truly large range. Broadly speaking, driving mobility has followed a similar pattern to what we saw for the whole country. Looking closer at Corpus Christi however, we see that it was the last city to drop in activity and the first to rebound — in fact, it has been indexing at over 100 since May 1. The remaining cities have been much slower to increase mobility, with Austin posting the lowest indices.
This kind of data is exactly why I am so excited about my new career. Data scientists at companies like Apple are able to do great things for the bottom line but they are also able to provide context for what is going on around us. I can’t wait to work on amazing projects like this one.