How can we make tourism more sustainable? The project “Data-driven Tourism for Sustainability” uses artificial intelligence and agent-based modeling to answer this question. In a conversation with Stefanie Wallinger from the Salzburg University of Applied Sciences, you will learn how digital technologies can simulate and optimize visitor flows. From the historic old town of Salzburg to the rural region of Bruck/Fusch, the project highlights innovative approaches.

Today, I would like to introduce you to an exciting project that is focused on the future of sustainable tourism: “Data-driven Tourism for Sustainability”. This project innovatively combines sustainability in tourism with data-driven decision-making.

To learn more about this initiative, I talked to Stefanie Wallinger from the Salzburg University of Applied Sciences. The university was one of the partners in this project, which uses artificial intelligence and agent-based modeling to develop sustainable solutions for tourism.

In this post, I share with you the key insights from our conversation. You will learn how this project could influence the future of sustainable tourism. For those who want to dig deeper into the topic, the official project website https://project-dts.eu/ provides more details.

Stefanie, can you give us an overview of the “Data-driven Tourism for Sustainability” project? What are the main objectives and who are the partners involved?
Stefanie: The project focused on the use of artificial intelligence and agent-based modeling for sustainable regional development using the example of visitor flows in tourist destinations over a period of three years (2021-2024). In addition to the lead partner, Danube University Krems (Department for E-Governance and Administration) with project manager Dr. Thomas Lampoltshammer, a team from Graz University of Technology supported the project as a further university partner. Datenvorsprung was on board as a technology partner with expertise in the field of AI, as was Nexyo as an expert in the field of data ecosystems.

The aim was to create a scalable data exchange platform on the one hand and to learn and simulate the authentic mobility behavior of tourists within two selected pilot regions using existing data on the other. The best-case scenario would be to model “what-if” scenarios that could help future decision-makers to make data-based decisions.

How did the idea for this project come about and which business partners did you have to implement it in practice?
Stefanie: We deliberately chose two regions with different levels of technological maturity as use cases. On the one hand, the city of Salzburg is a destination with a wealth of data. It is a city with a very high visitor frequency and a dense range of tourist attractions. On the other hand, Bruck/Fusch is a rural destination in the Hohe Tauern National Park that is often only passed through on the way to the Großglockner High Alpine Road.

This difference between the regions is also reflected in the quantity of data. In the city of Salzburg, the data from the SalzburgCard in particular provides insights into the visiting behavior of 30 tourist attractions. For example, it shows the order in which tourists visit various attractions in Salzburg, how long they stay at each attraction, and which days of the week they prefer to visit. This data is supplemented with information from the overnight stay statistics. For example, how many people stayed in which hotel and when. All other information comes from an AI model.

What data sources do you use for your simulations and how is this data collected?
Stefanie: Currently, the simulations only use historical data. However, the whole thing can be expanded infinitely. For example, weather data or mobile phone data would also be interesting to capture tourists without the SalzburgCard or to show the movement patterns of the local population.

Can you explain in more detail how artificial intelligence is used in your project?
Stefanie: An AI model condenses the information from the data basis and defines the specific route that each individual agent, i.e. tourist, covers within a day at the destination. From the information about hotel stays, attraction opening times, socio-demographic data, point of sale and the use of the SalzburgCard, a separate route is then defined for each point in the simulation, so to speak. The AI model makes a significant contribution to the authenticity of these movement patterns. Of course, these behavior patterns must also be validated accordingly. To do this, tourism experts from the respective pilot regions were consulted within the project, among others.

How does the AI help to simulate the movements of tourists and what advantages does this offer compared to traditional methods?
Stefanie: A clear advantage over traditional models is that agent-based simulations provide evidence-based information. As mentioned at the beginning, “what-if” scenarios can be simulated, for example, which can then be used as a basis for decisions. For example, if you want to open a new museum in a city, you could use the simulation to evaluate different locations by observing how the mobility behavior of the agents in the model changes. The main goal of the project was to demonstrate the possibilities of agent-based models in tourism by simulating visitor flows in two selected pilot regions. The more data you feed in, the more accurate it naturally becomes.

What types of algorithms and machine learning techniques do you use in your models? Perhaps you can briefly explain the modeling approaches and why you chose them.
Stefanie: We decided to use agent-based simulations (ABS). These always take place within a temporally and spatially limited space with a changing number of actors, whose behavior is defined by certain rules. Depending on the model, the agents can then, for example, coordinate their behavior, learn from each other and react to stimuli in their environment. An agent in the city of Salzburg could behave differently on a rainy day than on a sunny day and make visiting certain attractions dependent on the behavior of other agents or decide differently depending on age and gender. In addition, there is a certain random factor, because human behavior is not always comprehensible.

It would also be interesting, for example, to train the AI to recognize different types of tourists at a destination. This is probably where the strength of the method lies. Scenarios can be endlessly expanded depending on the rules the modeler adds.

How does your project contribute to promoting sustainable tourism?
Stefanie: The goals of the project are to be thought of in somewhat longer terms here. While the project is primarily intended to demonstrate the potential of data-based simulations, it can certainly contribute a great deal to sustainable regional development if it is continued. Therefore, a sustainability concept was developed within the project that deals with the scalability of the project results based on existing tourism strategies. In addition, our team at the Salzburg University of Applied Sciences, together with the Danube University Krems, conducted around 25 interviews with experts from the fields of tourism, data ecosystems, mobility, research and regional development.

Overall, agent-based models help to map the complexity of tourism systems and thus provide political decision-makers with a basis for mobility concepts, for example, from which both tourists and locals benefit. That is why we have also placed a strong focus on socio-cultural sustainability in the pilot region of Salzburg City. Especially in the context of mobility in city tourism, solutions are needed here that benefit both locals and guests.

Of course, the situation is somewhat different in rural areas. In the case of our second project region, Bruck/Fusch, the question arose at the beginning of the project as to how visitors to the busy High Alpine Road could also be attracted to Bruck/Fusch and how they could be motivated to stay longer. The aim of the simulation was, once again, to first show different movement patterns and to understand these using the model. In terms of sustainable mobility, it would of course be interesting to include data from different public transport or bus groups here.

To what extent can the results of your simulations influence the planning and management of tourist infrastructure to promote sustainability?
Stefanie: As already mentioned, the strength lies in the evidence-based nature of the simulations. As a basis for decision-making, models of this kind have great potential. In addition, the simulations could also be made available directly to tourists, for example via an app that provides information about heavy traffic at various attractions.

How do you ensure that the data you use is anonymized and compliant with data protection laws?
Stefanie: We only used historical data. All data was anonymized and no personal data was used.

What future developments and expansions are you planning for the project?
Stefanie: Since the data available in our rural pilot region was very limited, it would be exciting to focus on a larger rural destination such as a national park in a follow-up project. After the three years of the project, we now know what is possible, and it would of course be exciting to implement something specific.

How do you see the role of AI and ABM in future tourism research and planning?
Stefanie: Although ABM has great potential, it is still a rather unknown method at the moment. Of course, the situation is quite different when it comes to the topic of artificial intelligence. This topic is currently dominating every discussion, both at universities and in business. However, I see the biggest difficulty at the moment as being the lack of expertise. As our interviews on ABM have also shown, tourism destinations and businesses often lack the necessary know-how. For example, they may be sitting on a treasure trove of data that they themselves use only to a very limited extent. Projects like ours make a very significant contribution to actively linking tourism research with the economy. Especially in the field of AI, there are many exciting points of contact here.”

The “Data-driven Tourism for Sustainability” project impressively demonstrates how modern technologies and data-based approaches can make tourism more sustainable. The combination of artificial intelligence and agent-based modeling opens up new possibilities for destinations to better understand and manage visitor flows.

A big thank you goes to Stefanie Wallinger from the Salzburg University of Applied Sciences for her time and valuable insights into this forward-looking project. Her and her colleagues’ commitment is helping to make tourism fit for the challenges of tomorrow.

Projects like this remind us that sustainability and innovation can go hand in hand. They inspire us to continue looking for creative solutions that meet the needs of travelers as well as those of destinations and their residents.

I hope this insight into “Data-driven Tourism for Sustainability” has also inspired you and opened up new perspectives. Let’s work together on a sustainable future for tourism!

Thanks for reading and see you next time!