Google Maps for Tourism Monitoring and Analytics: The Case of Cultural Tourism, Sweden

A supply-side monitor visualizes Google places labelled by suppliers as ‘cultural’. For this purpose, 13,915 cultural place types were retrieved through Google Map’s API in 2019 and mapped to cultural place categories, like arts, cultural heritage, religious, and natural heritage. On the base of this data, the share and geographical distribution of cultural place categories for Sweden (Fig. 1) and for cultural place types for Swedish regions (Fig. 2) can be displayed interactively.

Fig. 1) Share and geographical distribution of cultural tourism places in Sweden

Fig. 2) Regional distribution of cultural tourism offer [e.g. Kronobergs]

By contrast, travelers’ experience outcomes are evaluated on the base of travelers’ feedback. For this aim, 353,960 text-tuples related to cultural place categories were automatically extracted and analyzed by sentiment analysis (lexicon-based), topic detection and topic clustering (Latent Dirichlet Allocation). By doing so, most popular as well as top-rated topic-clusters can be displayed for each cultural place type (Fig. 3).

Fig. 3) Topic clusters ranked by positive sentiments with place types

Interestingly, from a traveler perspective, the relatively smaller cultural supply elements, i.e. arts and cultural heritage, are evaluated more positively than larger supply elements, i.e. natural heritage and religious tourism, respectively. Finally, sentiment distribution can be shown for place categories for each region (Fig. 4).

Fig. 4) Sentiment distribution by region and category

We conclude that most relevant analysis perspectives for real-time tourism monitoring and analytics can be adequately supported by the inexpensive data source of Google Maps. For future research, we envisage to also include adjectives from feedback data for better grasping travelers’ complex cultural tourism experience. Finally, we aim at analyzing traveler’s spatial behavior and movement patterns by employing machine learning methods (association rule & sequential pattern mining). By so doing, we hope to be able to monitor the re-start of cultural tourism after COVID-19 pandemic in real-time.

This research is conducted by Prof. Matthias Fuchs together with Prof. Wolfram Höpken, Institute for Digital Transformation, Ravensburg University and Tobias Eberle, DoubleSlash, Friedrichshafen, Germany.