Benefits and pitfalls of using tweets to assess destination sentiment

Abstract

Purpose Destination monitoring is crucial to understand performance and identify key points of differentiation. Visitor satisfaction is an essential driver of destination performance. With the fast-growing volume of user-generated content through social media, it is now possible to tap into very large amounts of data provided by travellers as they share their experiences. Analysing these data for consumer sentiment has become attractive for destinations and companies. The idea of drawing on social media sentiment for satisfaction monitoring aligns well with the broader move towards smart destinations and real-time information processing. Thus, this paper aims to examine whether the electronic word of mouth originating from Twitter posts offers a useful source for assessing destination sentiment. Importantly, this research examines what caveats need to be considered when interpreting the findings. Design/methodology/approach This research focusses on a prominent tourist destination situated on Australia’s East Coast, the Gold Coast. Using a geographically informed filtering process, a collection of tweets posted from within the Gold Coast destination was created and analysed. Metadata were analysed to assess the population of Twitter users, and sentiment analysis, using the Valence Aware Dictionary for Sentiment Reasoning algorithm, was performed. Findings Twitter posts provide considerable information, including about who is visiting and what sentiment visitors and residents express when sending tweets from a destination. They also uncover some challenges, including the “noise” of Twitter data and the fact that users are not representative of the broader population, in particular for international visitors. Research limitations/implications This paper highlights limitations such as lack of representativeness of the Twitter data, positive bias and the generic nature of many tweets. Suggestions for how to improve the analysis and value of tweets as a data source are made. Practical implications This paper contributes to understanding the value of non-traditional data sources for destination monitoring, in particular by highlighting some of the pitfalls of using information sources, such as Twitter. Further research steps have been identified, especially with a view to improving target-specific sentiment scores and the future employment of big-data approaches that involve integrating multiple data sources for destination performance monitoring. Social implications The identification of cost-effective ways of measuring and monitoring guest satisfaction can lead to improvements in destination management. This in turn will enhance customer experience and possibly even resident satisfaction. The social benefits, especially at times of considerable visitation pressure, can be important. Originality/value The use of Twitter data for the monitoring of visitor sentiment at tourist destinations is novel, and the analysis presented here provides unique insights into the potential, but also the caveats, of developing new, smart systems for tourism.

Sentiment analysis in hospitality and tourism: a thematic and methodological review

“…The third most popular type of analysis was those related to comparison. Among these studies, analysis of variance (Becken et al, 2019;Fu et al, 2019;Gkritzali et al, 2018; Articles in this category used less traditional methods techniques. For example, Luo and Zhai (2017) visualized sentiment scores as time series data and assessed the changes through the glance of critical events that occurred in a destination.…”

Twitter users’ coping behaviors during the COVID-19 lockdown: an analysis of tweets using mixed methods

“…Here polarity score is of type float and ranges between [−1.0 and +1.0]. VADER takes into consideration emojis, slangs, emoticons, degree modifiers and capitalizations for score calculation (Hutto and Gilbert, 2014; Becken et al , 2019; Borg and Boldt, 2020; Moutidis and Williams, 2020). Only Tweets in the English language were captured.…”

The competitive productivity (CP) of tourism destinations: an integrative conceptual framework and a reflection on big data and analytics

“…More recently, the research stream revolving around smart tourism destinations and smart cities has underscored the importance of Big Data and analytics to effectively manage and market tourism destinations and to inform operations, services and innovation processes, at the destination level (e.g., Bakıcı et al, 2013;Batty, 2013;Becken et al, 2019;Gajdošík, 2019;Shao et al, 2017;Wise & Heidari, 2019;Xiang et al, 2015;Zeng et al, 2020). "Smart tourism" has been described as "a distinct step in the evolution of ICT in tourism in that the physical and governance dimensions of tourism are entering the digital playing field, new levels of intelligence are achieved in tourism systems" (Gretzel et al, 2015: 180) as the ways in which tourism experiences are created, consumed and shared are different.…”