An eTourism Research Project by: Angela Pagiri, Ruihong Liu, Barbara Prodinger, Fabian Wettinger, and Inna Milashevskaia
Most of our lives are greatly impacted by tourism. And not just because we have been studying it for three semesters. No, living in Salzburg we are surrounded by tourists (well, maybe not now, but you know what we mean), we work in tourism, we hear about it in the news and we spend our spare time travelling. Dealing with tourism so much, we couldn’t but wonder about how satisfied our guests really are. When they dine out in restaurants, do they get the experience they wished for? Do they all want the same, or does the lady from Tokyo pay attention to different things than the student from Barcelona? How are we going to know? That’s where our idea started forming. We realized that there are ways to find out. Even better, they are actually very accessible! Think of TripAdvisor. You can find thousands of reviews of restaurants in Salzburg there. If only one had the time to read them all. This is how we started to learn everything about getting hold of large amounts of data. How to source them from travel review portals, how to analyse them with machine learning and how to build a tool from it for everyday use in restaurants or at destination management organisations.
The question that had outlined the process of our research and study was the following: “Which aspects of the dining experience are important for visitors in the city of Salzburg according to their cultural backgrounds?”
Therefore, we as students and authors, took reviews from TripAdvisor to hand. Within these reviews tourists express their dining experiences, so-called “user-generated content”. A software called “Octoparse” allowed us to extract the data. For the analysis, the summer season 2019 (1st of May to 31st of October) has been taken into account. With the extracted data, machine learning came into place. A software named “Orange” allowed us to analyse the unstructured text document. With an open-source data visualization and topic modelling, distinct themes could be formed. The “GLOBE” cultural framework was then taken to group the review data into cultural/societal clusters. These were used for the analysis of the results and findings to explore the particular aspects that have been mentioned the most in each cultural cluster.
What our workflow looked like in Orange
The study demonstrated that overall, the most important aspects are (as we called them) “staff”, “food-menu items”, “value for money”, “restaurant physical appearance”, “food authenticity”, “overall service”, “menu offers”, “food quality”, “atmosphere” and “recommendation”. The priority of these aspects varies in the ten distinct cultural clusters by GLOBE. Such information is highly valuable for restaurant owners and other tourism providers as well as Destination Management Organisations. Now, their products and services can be adapted, customer service improved, and promotional texts, images and videos tailored to the different preferences with unprecedented precision. Further studies could be expanded nationally or internationally and other fields, such as accommodation providers or attraction managers could benefit from this research in similar ways.