Although it is only October, we are already preparing for next year’s travel season. We spent the last months collecting and updating all tourism data for next year and all our systems are now ready to start predicting crowds in 2020. Learning from 2019 and listening to user feedback, we have made some adjustments to our crowd predictions.
While predicting crowds for next year, we needed some data that is not yet available. Therefore, we now use cutting edge data prediction models to predict crowds for events that are yet to be announced. The same for school vacations. Even though some countries have not yet formally announced when school vacations will exactly be, we schedule them in anyway based on our previous predictions. As a result, our data is getting better and better. You can read more about how and why we do this below.
Crowd score changes in 2020
The first important change is that we’ve adjusted our crowd score because of you all. Although many travelers found our crowd predictions useful, the score wasn’t very helpful in distinguishing between extremely busy days and ‘normal peak season days’. A discussion on Tripadvisor, pointing out our weaknesses, made us see that we needed to make some adjustments. A user, whom we made anonymous for privacy reasons, pretty much said he wouldn’t use Avoid Crowds and we concluded he was probably right.
So we went to work and made some changes. Our crowd prediction score was great at telling you when it was peak season. But, like the person on Tripadvisor, most users already know which months will be busy. So we made sure our website will not only show you when it is peak season, but will also differentiate between regular peak season and those horrible super busy days. For example: when there are many cruise ships in town during peak season, our crowd predictions score will be very high (e.g. 95-100). If it is a regular peak season day, the number will still be high but not a score of 100.
More school vacations
Furthermore, we started looking well beyond the European borders to predict crowds in 2020. We have collected data about school vacations and public holidays in countries such as China, Japan, Brazil, India and many other countries. We decide on which countries to add by keeping a close eye on international travel trends.
As a result we have now taken into account more school vacations and public holidays than ever before. Not only do these school vacations help us to predict crowds, we also publish the information on our website for your reference. You can find information on school vacations and public holidays for all countries we’ve taken into account in our predictions on our website.
Modeling for unknown dates
Not all governments publish official school vacation dates far in advance. As a result, we officially don’t know when exactly each school will have a vacation. In general, most school vacations are now planned and published until the summer. After the summer of 2020, we look at historic data and international trends to predict when these school vacations will take place.
This is also the case for larger events such as marathons, gay prides or for example the Venice film festival. We are able to accurately predict when they will be organized, but are not completely sure until they are officially announced. We do however add our predicted dates in to our crowd predictions. You can recognize a prediction by the ‘dates tbd‘ in our overview of what’s going on that day. Once the official date has been announced we take away the model and add the official dates.
New destinations in 2020
In 2020 we will add more destinations in Europe as well as destinations in the rest of the world. In the coming months, more Caribbean destinations will be uploaded. Right at the moment that peak season in that part of the world is about to start. Meanwhile, we will also add more European destinations. We will focus on the Mediterranean as well as the area around the Baltic Sea.
Later on in 2020, we will start adding destinations in the United States. We weren’t able to predict for the US so far because we lacked the necessary understanding of American travel patterns. We are excited that we are getting more and more information that we can use to start modeling in that part of the world as well.
As we will spend the last months of 2019 preparing for 2020, we want to thank our users for helping us grow and make our predictions better. We loved your feedback and encourage you to keep telling us what works and what doesn’t. Because, although it is always nice to hear positive things, we learn most from feedback on what you think we should improve. That is how we can keep growing.