Rare and unpredictable events (such as terrorism, war, epidemics, and natural disasters) can have a huge impact on tourism dynamics. As small-probability events, those situations are not properly approached by statistical models, which usually perform very well in the presence of predictable patterns.

Therefore, high performance in tourism prediction is necessarily related to the capability to better model the uncertainty associated with rare events. In this paper, the authors adopt median absolute deviation (MAD) to identify outliers, which normally express abnormalities in a given system. By analyzing the preliminary results obtained, the authors point out the need for further efforts to empower the current models for tourism prediction; it is considered a kind of economic priority in the case of Florida--the object of this study--to maintain its status as a top tourism destination.

Unfortunately, the shortness of the paper (“extended abstract”) doesn’t allow any deepening and, generally speaking, any extended discussion. However, the paper is informative and sounds suitable for a very generic audience. Moreover, the analysis described is underpinned by a consistent dataset.

This contribution is a good starting point for a more extended analysis and is definitely a valuable asset for further research. In general, the work could also inspire others; indeed, the approach proposed is very generic and can be applied in contexts different from Florida. It leads to a number of interesting studies aimed at addressing generic or specific problems by taking advantage of the knowledge achieved; for instance, the capabilities of reactions (from a tourism perspective) of different countries or states after unpredictable events could be analyzed, and specific strategies to mitigate the impact of those events on tourism could be defined accordingly.