LAUSR

Abstract Knowledge of how destination marketing organisations (DMOs) use Twitter is still limited. This study aimed to assess how DMOs' Twitter activity affects hotel occupancy in short-break holidays. Key dimensions of Twitter that may affect hotel occupancy in tourist destinations were first identified. A longitudinal study using data for 10 Spanish DMOs was conducted to forecast hotel occupancy. Twitter application programming interfaces were used to gather data on tweets by DMOs and retweets and likes by users. Text mining was used to analyse the tweets by DMOs, differentiating between tweets related to events, attractions, socialisation, and marketing. Data were analysed using artificial neural networks. The best fit was achieved with a multilayer perceptron artificial neural network. Findings suggest that the number of retweets and replies by users and the number of event tweets, tourist attraction tweets, and retweets by DMOs can predict the hotel occupancy rate for a given destination.