The recent development of User Generated Content and Social Media platforms goes hand in hand with the increased availability of Big Data techniques and machine learning algorithms, having paved the way to collecting and analyzing great volumes of data. In this study we scan imagery data of Instagram, extracted from travelling-related posts, to analyze and provide a quantitative measurement of the destination image and of its dynamics.

With the help of a Visual Object Recognition tool, we convert imagery content into text for 860,000 travel-related pictures posted on Instagram in Summer 2019 in several European islands. The output for each destination and each point in time is a vector of labels’ frequencies on a very fine-grained scale. We then extend the metrics proposed by Arabadzhyan et al (2020), the IDDI – Index of Distance in Destination Image, to text and emoticons included in pictures’ captions and comments. This way, it is possible to investigate the affective dimension of the destination image in a more precise way than by analysing text only. Specifically, we determine the features of pictures more likely to be associated with positive or negative sentiment. Moreover, by merging the dataset with information about external conditions (like events organized in the destination, weather conditions or climate events) it is possible to determine how sentiments are triggered by specific contingency states, and estimate how they impact on the perceived quality of the tourism experience, thus helping implement effective adaptation policies that could mitigate negative effects on destination image.