Keywords

1 Introduction

The impact of online reviews, which are part of electronic word-of-mouth or eWOM, through user-generated content (UGC), has significance because of the large-scale accessibility and versatility of content across communities with diverse interests (Verma & Dewani, 2021), serving as indicators of the quality of tourism services (Yang et al., 2018). Such reviews have become an integral part of users’ decision-making process towards experiential goods (Chu et al., 2022), which have denoted an influencing factor in consumer behaviour (Cox et al., 2009), generating a bond of trust towards the experiences conveyed on review web platforms.

This research analyses the state of mind of TripAdvisor reviews, referring to the cultural tourist resources that configure the tourist offer of Malaga, which are ideal for the current study, through data mining, as they are unstructured texts. For this purpose, a qualitative analysis is carried out, examining the content, the themes of the reviews, and the feelings expressed by the users. By delving into the qualitative aspects of the reviews, this research aims to provide a deeper understanding of the impact of the reviews on the image of Málaga as a destination.

2 Literature Review and Research Background

2.1 Cultural Tourism in Málaga

The tourist destination of the municipality of Málaga, which is primary constructed on the “sun and beach” binomial, is one of the references that has positioned Málaga as one of the relevant points of the tourist offer of the Costa del Sol. However, over the last few decades, Malaga has undergone considerable progress in the cultural sphere, and the revaluation of heritage monuments, whose interventions have placed cultural tourism as the second reason for visits (Costa del Sol Málaga, 2021).

The search for diversification of opportunities in tourism development is accompanied by the initiatives established in the various tourism plans that Málaga has, the latest being the Málaga 2030 Strategic Plan, which shows the commitment to cultural tourism in the last decade, generating the development of other segments such as meetings and congresses (Fundación CIEDES, 2023). The presence of cultural spaces that form part of Malaga's tourist offer has placed it in third place among Spanish cities, corresponding to the cultural offer in 2021, with 53.6%, behind the cities of Madrid, with 88.9% and Barcelona, with 66.3% (Fundación Contemporánea, 2022).

2.2 TripAdvisor, Relevance of Reviews for the Tourism Sector and Sentiment Analysis

TripAdvisor is one of the social networks for travellers, containing more than 1 billion comments and reviews from almost 8 million businesses affiliated with the platform (TripAdvisor, 2021). This platform is considered one of the early adopters of Web 2.0, as its information and advice indexes are simultaneously built from accumulated opinions (Valdivia et al., 2019).

Experiences, transmitted through opinions and reviews shared by users of the TripAdvisor web platform, are part of online reputation, highlighting the connectivity between users without pre-existing associations (Baggio et al., 2008); whose only connections are based on sharing reviews of experiences. These interrelated connections, through online dialogue, are a factor that influences decision-making by other users (Rubio Gil et al., 2017); highlighting the values of credibility in the face of unbiased opinions and, in addition, reliability causing an increase in the prominence of options towards decision-making (Gavilan et al., 2018).

In a report by TripAdvisor, (2022), the users reported recognition of the value of reviews as descriptive and useful, as well as the accuracy of the content across a wide range of travel information. These qualities have granted trust to the qualitative content of the reviews, whose distinctiveness is part of the image of a collaborative digital tool that is useful for tourism (Minkwitz, 2018) and popular among tourism users and researchers (Taecharungroj & Mathayomchan, 2019).

2.3 Social Media Sentiment Analysis

The opinions issued on review web platforms reflect quality on a certain scale, being sufficient to capture sentiment (TripAdvisor, 2021); they are complemented by the sentiment of trust through indicators that provide veracity in the comments, ranging from personal information to actual images, locations, and dates of access to the website (Lee et al., 2018).Through qualitative interpretation, based on text mining or sentiment analysis, they can be developed into two types of studies: the word-level study, in which the analysis relies on the detection of keywords and lexical affinity (Poria et al., 2012), and the concept-level study, which considers the expressions coming from different words (Qazi et al., 2016).

Currently, there are several proposals in the current study and analysis of sentiment that cover the qualitative aspect, ranging from the search for keywords to verifying correlations between terms that can generate terminological guidelines (Delvizio, 2018). The aim of the conducted studies on sentiment analysis is to determine the emotions and personality of the wording on specific topics (Ray et al., 2021). Obtaining results through sentiment analysis has been useful in the formulation of various recommendations and predictions, such as the proposal made by Hu et al. (2017) formulating a multi-text summarisation technique to identify the most reiterative phrases in hotel reviews. Likewise, the results obtained can be oriented towards the determination of predictions, through the usefulness of the opinions and resorting to the classification technique (Lee et al., 2018). However, the authors Farhadloo et al. (2016) determine that one of the challenges in the qualitative study of reviews lies in understanding the aspects and determining a relationship between the feelings associated with the aspects found or the state of mind of tourists (Li et al., 2018).

3 Data and Methods

This study has considered the reviews transmitted on the TripAdvisor web platform, popularly accepted by tourists and managers (Valdivia al., 2017); and whose information has significant effects on decision-making by other users (Seok et al., 2020). For this purpose, online reviews have been analysed in Spanish using the phrase “cultural tourism in Málaga” on TripAdvisor.es to understand the perception of users towards the cultural tourism resources that form part of the tourism offer of the municipality of Málaga, from 2017 to 2022, compiling a total of 18.227 reviews. In terms of user ratings, 83% were positive (excellent and very good); 11% corresponded to neutral opinions (normal) and 3% to negative ratings (bad and very bad), demonstrating the high level of acceptance and satisfaction in the users’ experiences.

For the treatment of the extracted texts in a crude manner, far from any formal structure, the data structuring has been employed, as proposed by the authors Farhadloo et al. (2016), intending to obtain semi-structured data, following the interest of the analysis of the sentiments of the reviews on TripAdvisor.

The first step was to analyse the TripAdvisor ratings quantitatively, segmenting them into positive (excellent and very good), neutral (normal), and negative (bad and very bad) ratings according to the categories of the cultural tourism resources, as shown in Table 1. The findings show that most of the users’ opinions are positive, which indicates a high level of satisfaction in visiting these cultural resources.

Table 1 Percentage evaluation of the reviews according to the categories of cultural tourism resources in the municipality of Malaga
Full size table

Demonstration through visualisation of sentiment analysis data from reviews, as an important part of the analysis of the macro data, as observed by Cheng and Jin (2019), has been implemented through the KH Coder software, a programme that enables qualitative content analysis or text mining, employed in this research, responding correctly to tourism-related texts (Povilanskas et al., 2016; Jurkus et al., 2022).

4 Results

Based on word frequency analysis using the KH Coder software, as shown in Table 2, seeds of sentiments are visible, which are the nouns that place us in the positive or negative context of the image of cultural tourism resources. The positive terms with the highest number of frequencies are the following: good (1063), interesting (934), nice (859), best (566), great (548) and recommended (542); which convey positive and favourable feelings about the image of the resources analysed. In addition, these expressions reflect the level of satisfaction of visitors to the cultural spaces, with the comments being overwhelmingly positive, emphasising the positive feeling towards the museum spaces in a more proportional manner.

Table 2 The most common words for cultural tourism resources in the municipality of Malaga, found in TripAdvisor reviews
Full size table

However, terms also have been detected that attribute a negative sentiment status with less than 200 recurrences, referring to the terms full (147) and expensive (126); in direct relation to the categories of wineries and vineyards and churches and cathedrals. Another group of comments expressed the following terms in relation to the rest of the cultural tourism resources, such as bad (122), incomplete (29), poor (27) and, finally, the term scarce (26).

In Fig. 1, the co-occurrence network graph of the reviews is presented, emphasising the semantic aspect and the high degrees between the terms. The result, using the limit of 120 co-occurrences, has shown a relationship between 57 words (N), with a density of 0.075. Based on the co-occurrence study, 7 subgraphs could be detected, the first one being museums, linking the expressions *interesting, *great, *good, *art. Likewise, the terms *car, *collection, *building and *old, are associated with the Automobile and Fashion Museum, as it is one of the spaces with a high index of positive valuation and favourable reviews, giving favourable recognition to the permanent collection it hosts and the architectural value of the space, in relation to the old building of the Royal Tobacco Factory or locally known as “la tabacalera”.

Fig. 1
An illustration of the network of co-occurrences. It exhibits historical centers, historical monuments, museums, visiting the monuments, experience, and satisfaction with the experience. The subgraph ranges from 1 to 7. The coefficient ranges from 0.1 to 0.4.

Network of co-occurrences of reviews issued on TripAdvisor

Full size image

The second sub-graph corresponds to the image of historical monuments, which have been related to the feelings of *recommendable, *worthwhile, *best, *precious, *impressive; highlighting the places of Gibralfaro and Alcazaba, obtaining a higher degree of correlation coefficient with the terms *spectacular and *views. The terms theatre and Roman, appear as another sub-graph; however, they have a strong link of co-occurrence with the historical monuments, which is why they have been included in this segment.

Subsequently, the third sub-graph corresponds to the state of feelings of the most emblematic resources that compose the spaces of the historic centre, having a direct co-occurrence with the terms *historic, *street, *plaza; in addition, the Cathedral is related to the terms *palace and *nice. The term Picasso has a high rate of co-occurrence with *nice, which alludes to the Picasso Museum in Málaga, followed by the expressions *house and *square in relation to the geographical location of the Málaga artist's house-museum.

The fourth sub-graph corresponds to visits to monuments, where the term entrance ticket is highly correlated with the terms *opinion, *free, *Sunday, *time and *price. Such expressions demonstrate an interest in the time slot of free visits to museums and historical monuments.

The fifth sub-graph is related to the favourable comments on the satisfaction of visiting cultural tourism resources with the expressions of *staff, *rooms and *children. It is worth specifying the didactic factor of these resources in relation to the experience of children. Lastly, there are the terms *walk and *pleasant, which demonstrate the high rate of satisfaction on the part of the users.

Finally, a correspondence analysis was carried out, as shown in Fig. 2, to visualise the relationships between the dimensions of the reviews, through the experiences towards the cultural tourism resources of the municipality of Málaga, where the expressions that accompany the experience of museums, are characterised by *exhibition, *interesting, *great. As far as historical monuments are concerned, we can see that Alcazaba, the Roman theatre, the Castle and Gardens of Gibralfaro are related to the experiences of *views and *wonderful, *spectacular and *walking; in addition, there is a valuation in the correspondence between the terms of the city with the expression of *impressive.

Fig. 2
A bubble plot of dimension 2 versus dimension 1. Two clusters are plotted with frequencies from 500 to 1000 to 1500 to 2000.

Correspondence analysis between the experiences and the cultural tourism resources of the municipality of Málaga

Full size image

Likewise, about the historic centre, experiences are characterised by the terms *precious and *beautiful; and churches and cathedrals, together with the expressions *essential and *impressive. In reference to the words emitted towards the museum spaces in Málaga, they are: *art, *great, *collection, *work, *interesting, *great, *painter.

5 Conclusions

The study and sentiment analysis of the reviews transmitted on TripAdvisor towards the cultural tourism resources that form part of the tourism offer in the municipality of Málaga, based on a qualitative study, have shown the high approval and high level of satisfaction of visitors, whose reviews demonstrate that users tend to evaluate the experiences globally, confirming the conclusion of Valdivia et al. (2019).

The opinions reflected the interest in the monumental heritage assets distributed throughout the city of Málaga, expressing admiration and appreciation for the symbolic connection (historical), demonstrating the visitor's empathy towards the cultural heritage, and, in addition, appreciation for the state of conservation of these monuments. The reviews issued towards the museums recognise the variety of the collections they host, both permanent and itinerant, highlighting the works of local artists (such is the case of the Museum of Málaga and the Picasso Museum), adding the textual emphasis of “good and interesting exhibitions”. Likewise, there is a group of visitors who emphasise the architectural value of the museums as an added value of the tourist resource, as is the case of the Automobile and Fashion Museum, recognising good practices in the conservation of an industrial monument for museum purposes. Similarly, the comments show a high rate of appreciation for the environmental component (gardens), in complementarity with the monumental cultural heritage, reflecting a positive assessment of destinations committed to sustainability, through the protection of green areas. This conclusion ratifies the research of authors Nowacki and Niezgoda (2023); on visitors’ interest in nature, possibilities of contemplation and connection with the environment, and intangible values concerning nature tourism (Teles da Mota & Pickering, 2020).

There is also an emphasis on enquiries about the economic value of tickets for cultural tourism resources, confirming the research carried out by TripAdvisor (2021), which stipulates that 44% of visitors resort to the description of places and 36% to enquiries about the cost of tickets, as elements prior to trip planning, with the aim of making a more precise budget.

Another point to highlight is the recognition of the pedagogical factor in the management of cultural tourism monuments, which continue to disseminate content in a didactic manner for a diverse public, with appropriate information and, in the case of museums, with recreational activities for the attention of the younger public.

Overall, online reviews on TripAdvisor and other opinion platforms are relevant for a better understanding of the language of reviews and are indispensable for the tourism sector, as stated by authors Bridges and Vasquez (2018). Online reviews not only have an impact on the decision-making of other tourists but also influence how the entities in charge manage their offers. Similarly, reviews can provide insights into initiatives to promote authentic and enriching experiences, with the aim of preserving cultural heritage and encouraging sustainable practices.

The current study reveals several research topics. Firstly, comparative studies using machine learning. Secondly, using variables such as descriptions, the responsibility of the resource management bodies towards users and the sustainability of resources, with special attention to the divergences that exist between the information considered relevant by users and that prioritised by the TripAdvisor web platform. In addition, linguistic and semantic analyses of the use of different languages in tourism social networks.