Navigate back to the homepage

About Google's Mobility Data and COVID-19

Péter Péterdi
October 12th, 2020 · 3 min read

Google made its mobility data available as the situation around COVID-19 evolved and heard that these type of insights might be helpful for public health officials. Google Mobility data provides insights in 6 areas: grocery & pharmacy, parks, residential, retail & recreation, transit stations and workplaces. The data provided shows how visits and length of stay at different places changed compared to a predetermined baseline. The baseline for each area is the median value for the corresponding day of the week during a 5 week period from January 3rd to February 6 in 2020.

I must emphasize that in this article I don’t attempt to draw any conclusion between the popularity change of certain places and COVID-19 spread, instead, I will display how our lives and habits changed since COVID-19.


The Report

I created a Power BI report based on this data which you can discover below. For optimal viewing experience open it in full screen by clicking the icon in the bottom right corner.


Findings

In the next sections, I will only cover Hungary but in the report you may select other countries by opening the sidebar and choosing another country or sub-region.

Grocery & Pharmacy

At the beginning of March, a vast amount of news coverage mentioned that people started to stock up groceries, however, in Hungary, this clearly was not the case. As news spread, there was a slight increment in this area, but later it evened out and had significant decreases as well, particularly on holidays. It’s a common thing in Hungary (and surely in other parts of the world as well) that around a national holiday people overrun the grocery stores, which happened around March 15 as well, followed by a huge decrease in popularity in this area compared to baseline.

Parks

In March, people visited parks much more often than within the benchmark period and this trend was continuing as the temperature rose and summer came. At the end of March for the period of a brief week between March 21 and 26 parks’ popularity decreased, but so did every other area’s apart from residential. Hungary passed 100 active cases on March 22.

Residential

Until March 15, there was no significant change to baseline in the popularity of residential areas. After March 15, this area increased by an average of 20%, although the volume of increase has steadily become smaller. It is also worth mentioning that each week’s Saturdays and Sundays had the smallest change to baseline as most people were also at home in the benchmark period and weekends had the slightest change in our lives in these times. As June came weekends even had a negative change to baseline, supposedly as people were trying to relax and do something fun outdoor as we can see it in the increased popularity of parks in this time frame.

Retail & Recreation

From March 15 to May 29, we had a two and a half month period when the popularity of this area was negative compared to baseline and only with the start of the summer people began to spend more time in retails. This two and a half month period may be explained by the cautiousness of people in this uncertain time. It’s too worth pointing out that typically Sundays had the largest change to baseline in the summer - as could be predicted - frequently reaching 50 percent.

Transit Stations

Transit stations are heavily correlating with the Retail & Recreation category which may be explained by public transport being the most popular way of transportation. On weekdays, transit stations’ popularity had been increasing, but stayed still below the baseline numbers, until mid-July. The largest drop in this category was after March 15, which implies that more people started to utilize alternate means of transport or had to work from home, as seen in the data about residential areas and workplaces.

Workplaces

As mentioned in the transit stations’ section, after March 15 workplaces were much less visited compared to the benchmark period, and even at the time of writing this article it is down by an average of 20 percent on weekdays. On weekends, the change is not that significant, naturally.

Holidays

When analyzing the data, most of the record high changes can be explained by a holiday or a working Saturday. March 15 resulted in a significant increase in the popularity of parks and a significant decrease in the grocery & pharmacy category. At Easter and May 1, most people were at home and every other category apart from parks had a significant decrease. June 1 (Pentecost) showed very similar data to Easter and May 1, except parks were extremely popular on this day. August 20 also had very similar changes like the other holidays, while August 29 had a record increase in workplaces category by +41 percent compared to baseline as it was a working Saturday.

More articles from Péter Péterdi

Who Are The Top Players of NBA According to Data?

Paul Pierce - former NBA champion with Celtics and MVP in 2008 - said recently, that NBA's current best, LeBron James isn't part of the 5 all-time bests. Let me clarify. He claims that Michael Jordan, Bill Russel, Kareem Abdul-Jabbar, Magic Johnson, Tim Duncan, Kobe Bryant, and Larry Bird are all better players than LeBron. Of course, James and Pierce have history, they never liked one another, however, this statement and Netflix's recent big hit, The Last Dance gave me the idea to delve into this topic and examine who are the best players of NBA according to data.

July 4th, 2020 · 4 min read

Why I started this blog?

My take on using data in 2020.

July 3rd, 2020 · 1 min read
© 2020–2021 Péter Péterdi
Link to $https://www.linkedin.com/in/p%C3%A9ter-p%C3%A9terdi-09896768/Link to $https://github.com/pterdi