Topics: Coronavirus | AI | Data Science | Deep Learning | Machine Learning | Python | R | Statistics

KDnuggets Home » News » 2020 » Aug » Tutorials, Overviews » Visualizing the Mobility Trends in European Countries Affected by COVID-19 ( 20:n32 )

Visualizing the Mobility Trends in European Countries Affected by COVID-19


This post highlights the movement of people from the 10 most-affected European countries based on the way they stay at home, work, and visit places, using Google's anonymized location tracking dataset.



By Preetish Panda, Marketing at Tribe

COVID-19 is going to draw a distinct mark in the timeline of the world -- one before and one after. It has completely transformed the world within a span of a few months. And one of the key changes it has brought about is mobility, i.e., the way the people move.

Primarily because of the way different governments across the world have implemented “lockdown” policies and actively encouraging people to stay at home. This method of enforcing physical distancing has reduced the spread of the coronavirus. 

In this post, I’ll highlight the movement of people from the 10 most-affected European countries based on the way they stay at home, work, and visit places. Since Google is at the top of its game when it comes to tracking people, the search giant leveraged location tracker to create an anonymized dataset. 

This dataset is called COVID-19 Community Mobility Report and it has the following location categories:

  • Grocery and pharmacy
  • Parks
  • Transit stations
  • Retail and recreation
  • Residential
  • Workplaces

According to Google, the value for the location categories shows the changes from the baseline value for that day of the week. And the baseline is the median value, for the corresponding day of the week, during the previous 5 weeks (Jan 3–Feb 6, 2020).

 

Preparing the dataset

 
The dataset contains the date, country, and region fields along with the corresponding value for the location categories. Here is a screenshot of the partial tabular presentation of the dataset:

Image

 

For this analysis, I removed the sub-region fields and selected the country-level data. Further, I narrowed it down to the top 10 most-affected European countries (as of 30th June 2020). These countries are - Italy, France, Germany, the UK, Belgium, Spain, Portugal, Netherlands, Sweden, and Switzerland.

 

Visualizing non-residential movement

 
Let’s now look at the way people have been moving, in non-residential areas by creating time-series charts with a 7-day moving average for each country.

Italy

We can see that from the middle of March to the beginning of May, there has been decreased mobility. Lately, towards the end of May visits to parks have increased in comparison to other locations. This is mostly due to the ease in restrictions imposed by the government in May.

Image

 

Sweden

Sweden has largely relied on its citizens to take the right decision. The movement has remained lower in all the location types, except Parks. Towards the end of May, the movements in Parks have considerably spiked and the trend continues in July as well.

Learn more about Sweden’s controversial COVID management strategy here.

Image

 

Germany

The movement patterns in Germany is similar to Sweden (notice the plot for the parks?). This is also related to the ease in lockdown and the start of phased reopening. Mobility in “Grocery and pharmacy” locations has remained much closer to the baseline. 

Image

 

Spain

In March, we can see there was a sharp fall in movement. Retail and recreation locations are the ones with the least mobility and parks are experiencing a spike in movement.

Image

 

Belgium 

Visits to parks are going through the roof now. Understandably, movements related to “Grocery and pharmacy” are reaching the baseline.

Image

 

Portugal

In February there were increased visits to parks and it reduced over the course subsequent months. During May, it has increased and crossed the baseline value.

Image

 

France

There was one spike in visits to the “Grocery and pharmacy” location in February and eventually dropped throughout March and April. Finally, it substantially increased in May. We can also see a spike in “Grocery and pharmacy” visits in March -- perhaps people were preparing for the lockdown by purchasing essential items.

Image

 

Switzerland

In Switzerland, right from February, parks are experiencing high levels of movement from the baseline. Also, “Retail and recreation” locations are the ones with the least mobility.

Image

 

The United Kingdom

In the UK, although people reduced the visits to the parks from March, during May it has spiked owing to lesser restrictions. From May to July the visits to parks has fluctuated with a peak in July.

Image

 

Netherlands

Clearly, during the last few months, the movements in the parks have been more than the baseline value because of the phased reopening in May. And transits are the location categories with the least movement.

Image

 

 

Visualizing residential movement

 
We looked at the non-residential movement, now let’s see how people have been staying at home over the past few months. Overall, the general trend remains the same across the countries -- people started moving out of their homes during May. 

The trend continues in July as well except for Sweden. In Sweden, from July people reversed the residential mobility and started staying at home. It could be because of the increase in COVID infections towards the end of June in Sweden.

The UK is another country that stands out in terms of residential mobility trends. During May, the movements did not drop sharply in comparison to the previous months and there has been a slower change through June and July. 

Image

 

Let’s now look into the data for February and July for all the countries. The dumbbell chart given below shows the change in average residential mobility for the two months. Although we should not compare countries as the lockdown measures widely differ based on the locations, we see that presently the people from the UK and Portugal are the ones who are at the top comes to staying at home.

In the case of Italy, the residential movement is back to the February level.

Image

 

 

Moving to the future

 
We can already see that the countries have loosened the restrictions and people are increasingly visiting non-residential locations. What would be the impact of that? Thanks to data, we don’t have to wait longer!

This type of anonymized datasets can immensely help in predicting localized outbreaks, planning public transits, and above all, understanding how people behave. No wonder, the European Commission is quite inclined to work with Europe’s telcos to access aggregate location data of their subscribers.

However, the legality of tracking location data, especially the way companies get consent from people remains a dubious affair. Perhaps when the world is in greater danger, there is a trade-off that needs to be arranged between people and their governments.

 
Bio: Preetish Panda handles marketing at Tribe, a platform to build branded online communities. Professionally he is passionate about brand building, insight mining, web, and mobile technologies.

Related:


Sign Up

By subscribing you accept KDnuggets Privacy Policy