COVID-19 Community Mobility Reports from Google and Apple

Google and Apple manage a large swath of global mobile data. Whenever a iOS or Android user connects with a map or search engine a geo-coordinate trail is created. A location footprint is left by the mobile device IP address with a latitude and longitude of the location and a timestamp.

With the urgent need to better understand the spread of COVID-19 many mobile device companies are making this location data available. We must applaud Google and Apple for not only making that data available to everybody in an easy comma-delimited file but also to ensure it’s anonymized.

Let’s take a quick look at the cool things you can chart and learn from these global datasets across countries, regions, and big cities.

 

Google Mobility Data

Google offers a series of categories in the mobility report including:

  • Retail & Recreation
  • Grocery & Pharmacy
  • Parks
  • Transit Stations
  • Workplaces 
  • Residential

Lombardy, Italy

Let’s explore the region of Lombardy, Italy with this dataset. This was the hardest-hit area in Europe right after Wuhan, China. The order to stay at home was given March 10, 2020. It’s interesting to see that a drop in retail activity and a rise in residential activity was seen as early as February 22, 2020. 

The zero line represents the baseline for “Retail and Recreation”, the Lombardy region was way below its normal levels a month before the official stay at home order.

Also of interest is that people left their homes on weekends, as seen with the dips in the residential chart.

Comparing US States

Let’s see who leads in terms of stay-at-home behavior in the United States. If we pull the “Grocery and Pharmacy” category and calculate the overall mean for all states we see Hawaii with the lowest numbers, meaning they are way below their normal baseline for groceries. Iowa, on the other hand, changed its normal behavior but did not drop below the baseline level (and an interesting peak upwards in mid-March, panic buying?).

Comparing Behavior with Google Trends

Just for fun, let’s bring in Google Trends. Let’s look up the term “Hawaii Parks” and download the Google search-engine activity from Google Trends. When we plot it along with the “Parks” category of the Google COVID-19 Community Mobility Report, we see a surge of search activity on March 16, 2020 along with a drop in mobility activity. Clearly, folks were curious about what they should do regarding park activities a week before the official stay-at-home order went into effect on March 21, 2020. 

Apple Mobility Report

Apple offers a mobility report based on Apple Maps utilization. The landing page for the report states that “privacy is one of our core values, so Maps doesn’t associate your data with your Apple ID, and Apple doesn’t keep a history of where you’ve been”. Hopefully, this should assuage privacy concerns. 

Their mobility report offers three “Transportation Categories”:

  • Driving
  • Transit
  • Walking

Let’s take a look at New York City’s transportation behavior. The stay-at-home order was given on March 22nd, 2020. All three categories show similar behavior, they started showing changes as early as March 10th, 2020. Transit flattened out after the stay-at-home order but city walking and driving still saw spikes in activity on weekends. 

What’s Going On?

I’ll end this mobility blog with a chart from Barcelona. Notice that huge peak prior to the stay-at-home order? In late February Barcelona residents enjoy a series of back-to-back holidays. The town opens up buildings, parks are full of activities, and people use the transit system to catch the festivities. This is normally a fun period in Barcelona, but this year, it may have helped the spread of Coronavirus.

If you are interested in finding out more about how to use these mobility reports, feel free to reach out to us at info@springml.com or attend one of our webinars.