While tourism plays an important role in the growth of the global economy, tourism practitioners still face challenges in understanding the complex travel patterns of tourists. Travel diaries based on the travel history recorded by tourists during their trips, especially sequential association of the visited locations, can effectively reflect their travel behaviors and preferences. For instance, tourists who visited Canada would also tend to visit the United States during the same trip. Such sequential associations are helpful for travel agencies in formulate more appropriate business strategies and travel products to meet tourists’ need.
In view of this, Prof. Rob Law of the School of Hotel and Tourism Management and his co-researchers developed a novel data mining technique to extract tourists’ travel patterns from photos that are uploaded to the photo-sharing website Flickr. Firstly, the researchers collected 809,313 publicly available photos uploaded by 3,623 Flickr users between 2001 and 2015. Subsequently, based on the photos’ Global Positioning System (GPS) information, they were converted to sequences of visited destinations where information like countries and cities visited were extracted. Utilising this sequential rules mining method, photos were sorted and processed systematically according to location and time.
The researchers obtained a total of 17,188 travel diaries and were able to discover tourists’ travel patterns and behavior such as the number of countries visited on one trip, the most popular destinations etc. Further application of this novel data-mining technique would offer the possibility of analysing data at more in depth and micro-levels, such as individual tourist sites visited, travel styles, preferences and purposes etc.