Impact of COVID-19 on tourism

Impact of COVID-19 on tourism main image

Context and motivation

The outbreak of Covid-19 had a major impact on our everyday lives. In just a couple of weeks, all our habits were shaken: we started to wear masks, work from home, and maintain a safe distance from each other. These tremendous changes inevitably reflected on our economy: with people cooped up at home and the borders locking down, most commercial activities had to be temporarily suspended; some were even shut down permanently. The main question on everyone’s lips went from “When will the situation be back to normal?” to “Will the situation get back to normal?”. This study is an attempt to answer this question.

As the Covid-19 pandemic affects so many lives, it has been imperative for governments, individuals and businesses to track the way it has affected society. They need to foresee the consequences of the disease and lockdown measures and adapt accordingly. Tourism is an important component in the global economy and one of the most impacted sectors. In some regions of the globe, tourism is the dominant component of the local economy. We show in this blog that not only tourism economy has been massively impacted by the virus outbreak, but also the tourism destinations are changing.

Our method

At a high-level, our goal is to define as many “indicators” as we can to closely monitor society. An indicator can be virtually any metric, provided that we can measure it in a quantifiable manner over time. Since the Covid-19 outbreak has roughly changed every aspect of the society, any such indicator will likely see its trend or pattern scrambled around mid of March, when Covid-19 got viral and propagated all over Europe.

This blog focuses on tourism. The indicators of interest are the number and location of Airbnb properties booked over time. First, we look at how the number of bookings and we compare these numbers before and after the COVID outbreak. Not surprisingly, the pandemic has largely reduced the number of travellers and this is reflected in the number of bookings that dropped significantly as soon as the countries went into lockdown. Next, we looked at the destinations chosen by tourists after the outbreak. The assumption was that tourists would rather visit remote and less crowded places instead of densely populated areas.

Code Available on Github

Tourism in the UK

In many regions of the world, tourism is a seasonal and periodic process. The picture below shows the number of properties booked over time in Bristol, UK. We can see that tourism was growing exponentially in that region but also it has a clear seasonal component to it. It goes up in summer and down in winter.

A chart showing the number of properties booked over time in Bristol, UK

To better see the trend and period behaviour, we used the Facebook prophet open-source library to decompose the time-series shown above into its trend (top) and its yearly seasonality (bottom).

A chart showing the number of properties booked over time in Bristol, UK with help from Facebook prophet

The same growth and seasonality have been observed in different cities in the UK, namely London and Edinburgh. The table below summarises the growth in the last six months, as well as the average monthly growth in these three regions.

A tablesummarising the growth in the last six months, as well as the average monthly growth in London, Edinburgh and Bristol

Note that these measures are computed based on the data downloaded from Inside Airbnb. Since the actual reservation dates of the properties are not available, we used the date of the reviews instead (knowing that a user can only leave a comment after he has rent a property). The dates of the reservations and reviews might slightly differ since in most cases, users will give a review only when their trip is over. In the future analysis, the team could consider shifting the date of the reviews (e.g. by subtracting 5-10 days) to estimate more accurately the check-in day of the user.

Without any surprise, we also see from the pictures above that the number of reservations dropped from the beginning of 2020, as soon as COVID hit the country and the European continent. In the next section, we will set our focus on the small number of reservations made after the end of the first quarter of 2020. We will show that, besides the crash of the number of reservations, something else happened: travellers are changing their favourite destinations from hot, touristic, crowded places for remote and less populated areas.

Geo-Distribution of tourism within UK

Even if the demand for Airbnb dropped dramatically due to the Covid-19 situation, tourism did not stop. People are still travelling. Looking at the geographical distribution of the number of Airbnb reviews over time offers a couple of thought-provoking insights into people’s behaviour.

As showed above, before Covid-19 crisis the trend in tourism has rapidly increased. However, people were usually choosing properties in the city centre, close to touristic attractions and lively places. When the lockdown started in the UK at the end of March, people’s choices started modifying.

Taking the number of reviews as an indicator for Airbnb demand and people’s mobility, we normalised the monthly data and obtained a geographical distribution of the most appealing neighbourhoods in three cities in the UK (Bristol, Edinburgh, and London). This method discards the effect of the drop in the number of reviews. In all three cases, the distribution starts showing a new trend in travelling just after the lockdown started in April 2020. We can see that an equivalent percentage of travellers who, before the pandemic, preferred the city centre as holiday accommodation are now choosing more remote areas. The effect of this behaviour may be a result of the fear of crowded spaces or it may show a lower budget for holidays. Due to incomplete data for May or June, we cannot make any further assessment.

However, where are these travellers coming from? Looking at the lockdown measures in the UK in April, the UK government advised against all non-essential travel. Taking this into account, it can be implied that tourism happens locally. In other words, the decrease of international flights and the present Airbnb travellers indicate local tourism.

Graph charts showing the number of reviews distribution each month in Edinburgh

Fig 3: Edinburgh Neighbourhoods – Number of reviews distribution per month

Graph charts showing the number of reviews distribution each month in Bristol

Fig 4: Bristol Neighbourhoods – Number of reviews distribution per month

Graph charts showing the number of reviews distribution each month in London

Fig 5: London Neighbourhoods – Number of reviews distribution per month

Main Takeaways

For many countries around the world, tourism has been one of the major sources of revenue. Both international and local travellers have given tremendous financial output to countries’ GDP. On one side, the pandemic has hit hard various industry sectors. On the other side, it has been changing our behaviours and challenging the way we work, travel and react to a global crisis. Understanding the shift in thinking and the evolving societal trends will drive the industries back into revival and will define the New Normal.

Considering the above results, the tourism industry and people’s mobility are changing rapidly from one month to the next. What was conventional a year ago is becoming out-of-date nowadays. After analysing the data the following trends were observed:

  • prior to the pandemic, the demand in Airbnb was increasing as fast as 10% over six months with yearly seasonality;
  • during the pandemic the demand for Airbnb has not dropped to zero;
  • people are preferring more remote areas as holiday destinations;
  • local tourism escalated over international tourism.

Disclaimer: This information can be used for educational and research use. Please note that this analysis is made on a subset of available data. The authors do not recommend generalising the results and conclude decision-making on these sources only.


Vincent Nelis is Senior Data Scientist with IBM Data Science & AI Elite team where he specialises in Data Science, Analytics platforms, and Machine Learning solutions.

Maria Ivanciu is AI Developer with Rolls-Royce R2 Data Labs team where she specialises in Data Science and Machine Learning solutions.

Special thanks to Erika Agostinelli, Klaus Paul and Mehrnoosh Vahdat who helped us in this work.