Google data analysis

Since April 3th, Google has been releasing so called 'Mobility Reports' in which they show for each country to what extent their users are visiting certain types of places (e.g. restaurants or public transport hubs). With over 2500 million people having a Google device, the data is extremely valuable and worth analyzing. In this analysis I use the Google Mobility Reports of eight different European countries to see to what extent these countries have been 'social distancing'.

Methodology

Google Mobility Reports track the millions of identified places (restaurants, libraries, grocery stores, etc.) by aggregating these into six categories:

1. Retail and recreation (incl. restaurants, cafes, movie theaters, shopping centers, libraries)
2. Grocery and pharmacy (incl. grocery markets, super markets, food warehouses, drug stores)
3. Parks (incl. beaches, national parks, public gardens)
4. Transit stations (incl. subway, bus and train stations)
5. Workplaces
6. Residential

For the below comparison, I decided to use the category 'Retail and Recreation' as this category includes the places which were first targetted by governments who wanted to contain COVID-19 outbreak. The Google source data (1 March 2020 - 2 May 2020) however turned out to be rather volatile, as visualized below:

For readability purposes I calculated a 7 day moving average for each day and each country and visualized this chart for further analysis below:

Interpretation of the chart

The chart shows that the Google data confirms what one would have expected based upon the government policy decisions in recent months. Some examples:

• Countries which were in strict lockdown (e.g. Spain and Italy) saw a bigger reduction in mobility than countries which weren't (Sweden and Belarus).
• Italy was the first country in Europe to impose a lockdown, and Google data saw mobility reduction first going down in Italy as well (see above chart between March 4 - March 12).
• Austria was one of the first European countries to start reopening their country on April 14th, this is also clearly visible in the chart.

However there are also some surprising takeaways from the Google data:

• The Netherlands initially claimed lockdowns were just symbolic measures and would not be useful, but the government eventually introduced a 'intelligent lockdown' in March. The mobility reduction in the Netherlands is however surprisingly similar to countries without a lockdown, like Sweden and Belarus.
• In early March, Belarus was the country with the lowest reduction in mobility, however the amount of people visiting restaurants, cafes, etc. has significantly dropped in Belarus since. In fact: for most of April the mobility reduction in Belarus was higher than in Sweden. Looking to the original source data for May 2, the mobility reduction in Belarus is now almost equal to that of the Netherlands.

Source data: https://www.google.com/covid19/mobility/