CONICET | Buscador de Institutos y Recursos Humanos

Tourists are increasingly utilizing information from digitalized comments through tourist-generated content (Bigné, Ruiz & Currás, 2019). Travel review sites such as TripAdvisor relies upon tourist-generated media or electronic word-of-mouth (eWOM), providing tourists outstanding opportunities to share their opinions of travel and hospitality service products and brands with other consumers (Litvin, Goldsmith & Pan, 2018). eWOM through travel review sites influences travel related decisions and shapes perceptions and images about destinations and travel offerings. Within travel review sites, tourists are able to reflect openly about their holidays or seek travel information. eWOM websites provides researchers with an unprecedented opportunity to better understand tourist experiences (Mkono & Markwell, 2014). Learning from the big data organically created about the tourism destinations is the new challenge for tourism destinations managers.When a new paradigm shift is knocking on the door of marketing researchers (Bigné, 2016), text mining offers an answer for the question about dealing with unstructured text data. Text mining can be classified into linguistic and non-linguistic approaches (Villarroel, Theodoulidis, Burton, Gruber, & Zaki, 2014). Based on a corpus of 5.113 TripAdvisor online reviews, about activities and attractions of the ten best ranked destinations of Argentina and Spain, we examined and compared the eWOM of both countries complementing the strengths of both approaches. First, throughout non-linguistic techniques using Rapid Miner software, we analysed the reviews as a list of terms, counting the number of times that words appear in a document. This analysis enabled us to identify and compare the most frequent words and n-grams used by tourists when ?talking? about destinations. For instance, terms with extremely high frequency were: lugar (place), ciudad (city), bien (well), pena (worth), experience (experience), puede (can), recommendable (recommended), hacer (make/do), visita (visit), mejor (better) and paseo (walk), among others.Second, as linguistic techniques consider the natural language characteristics of the text in the documents (e.g., syntax, grammar), we automatically analysed the whole texts with Meaning Cloud software and identified their topic, subjectivity level and valence. The most frequent thematic arguments of the online reviews (eWOM of tourist destinations) were: ?art and culture? (21.9%), ?tourism, travel and commuting? (16.4%), ?environment, weather and energy? (14.1%) and ?lifestyle and leisure? (8.9%). We combine these results and run again the RapidMiner Studio analysis to identify the most frequent words in each topic category. For example, in the he comments about ?art and culture? the most mentioned words were "museum" (275 times), "city" (225 times), "place" (224 times), "visit" (213 times) and "cathedral" (193 times). While in those of the category ?environment, weather and energy? the most frequent words were lake (231 times) and place (217 times), followed by promenade, nature, glacier, park, landscape, forest, water, road, among others.Finally, we carried out a correspondence analysis that showed the relationships between destinations and topics categories. In nature destinations as Villa La Angostura, Bariloche, San Martín de los Andes, Calafate, Ushuaia e Iguazú tourist-generated content were more frequently related to the environment; while in cultural and urban destinations, as Salamanca, Madrid, Sevilla, Granada, Mar del Plata, Palma de Mallorca, Málaga, Salta, Barcelona y Buenos Aires their reviews were more related to education, art and culture, and religion.