Advancing the visual turn in tourism, hospitality and leisure studies Open Access

Visuals, in the form of both still and moving images, are becoming increasingly dominant in tourism, hospitality and leisure practice and research. The often-visual nature of representation, imagination, experience and consumption in these contexts therefore cannot be studied without the consideration, as Banks (2007) notes for the social world, of the visual. While the power of visuals has been widely acknowledged by tourism scholars (Urry and Larsen, 2011; Feighey, 2003; Rakić and Chambers, 2012a), the tourism, hospitality and leisure research landscape has historically been dominated by number- and word-based research. Nevertheless, as visual methodologies started to (re-)emerge initially in the wider social sciences and humanities (e.g. see Banks, 2001, 2007; Margolis and Pauwels, 2011; Pink, 2021; Rose, 2023) these developments were mirrored in tourism, hospitality and leisure studies (e.g. see Scarles, 2010; Rakić and Chambers, 2012a; Matteucci, 2013; Pernecky and Rakić, 2019; among other).

That said, it was only the subsequent proliferation and widespread establishment of these methods as legitimate tools of scholarly enquiry (Chambers, 2012; Rakić, 2012; Pernecky and Rakić, 2019) that brought the multidisciplinary fields of tourism, hospitality and leisure studies to a “visual turn”. This shift is evident not only in the proliferation of the variety of different visual data types, that are increasingly relied on in these fields, but also in the popularity of the visual within different, and at times also incommensurable between them, philosophical paradigms or positions. Namely, tourism, hospitality and leisure studies are currently witnessing a reliance on visuals among both studies underpinned by the “constructivist/constructionist” worldviews as well as studies grounded in the “positivist” worldviews (for a deeper discussion on these two core groups of worldviews or paradigms, and the incommensurably between them, see Lincoln et al., 2018, 2024).

A wide variety of visual data in these fields of studies to date most commonly relied on still images in the form of photographs produced by tourists, professional photographers, images and catalogues commissioned by private or public entities, and those produced by either researchers or research participants (Rakić and Chambers, 2012a; Volo and Irimias, 2021). In addition to photographs, still images relied on as visual data also included drawings (Hunter, 2012; Köstinger and Matteucci, 2024), paintings (Tribe, 2008; Staszak, 2012) and posters (Wiliams, 2012; Sel and Aktas, 2019). Less commonly, but still contributing to the overall visual turn, were moving images in the form of researcher-produced video including videography (Rakić, 2010; Hietanen, 2012; Masset et al., 2024; Kawaf and Dekel-Dachs, 2024; Burns, 2024; Relton, 2024; Akwasi Adu-Ampong, 2025) or moving images such as promotional videos relied on as data for the purpose of either elicitation (e.g. Shani et al., 2009) or analysis in research projects (Moin et al., 2020). Simultaneously, user-generated visual data sourced from secondary sources, such as internet memes (Yhee et al., 2024) are afforded less attention, although these have a potential as visual data which can be either analysed or relied on as a part of elicitation techniques.

In the context of efforts and calls for well underpinned incorporation of visual methods in research projects to continue (Volo and Irimias, 2021), especially given that societal, cultural, political, environmental and economic changes are overwhelmingly represented in a visual format – whether through social or traditional media – there is an urgent need to further contribute to advancing the visual turn in tourism, hospitality and leisure studies. These contributions can play their own unique role in (a) providing a greater balance within the methodological landscape between visual- text- and number-based methods and approaches, (b) advancing the reliance on visual methods which are well suited not only for the research projects within which these feature but also for their underlying research paradigms and (c) broadening the scope of the visual studies debate in tourism, hospitality and leisure research.

In this vein, the aim of this paper is threefold. This paper aims to (1) inspire further advancement of both the scope and breadth of visual research in tourism, hospitality and leisure; (2) offer an updated overview of the contemporary research philosophies or paradigms and their implications for visual data and (3) provide an insight into recent empirical investigations that have relied on either big or small visual data. In its conclusions, this paper also offers some pointers for future research with critical issues and debates in visual tourism, hospitality and leisure research.

To set a common framework on which to reflect, this paper offers an updated version of the philosophies of visual Table originally published by Rakić (2012), visually representing key paradigms and their implications for visual data while incorporating key developments in the philosophies of sciences which have taken place since. While acknowledging that it is not the aim of this paper to provide a detailed overview of each of the key paradigms (for that, refer to Lincoln et al., 2018 and Lincoln et al., 2024), and, that a variety of different paradigms will be emerging over time, this paper highlights two core groups of paradigms, namely, “positivist” and “constructivist/constructionist” worldviews, both more broadly speaking as well as in as far as their implications on visual data are concerned.

The differing implications for visual data, especially along the broad lines of groups of philosophical positions which are perceived as incommensurable between one another (e.g. see the division between “positivist” and “constructivist/constructionist” worldviews in the table above and Lincoln et al., 2018) are particularly relevant. As Lincoln et al. highlighted commensurability becomes ‘an issue only when researchers want to “pick and choose” among the axioms of positivists and interpretivist models because the axioms are contradictory and mutually exclusive” (2018, p. 133). Illustrating this point in the context of visual data in the table above, is the fact that visual data such as a photograph cannot be simultaneously perceived as both a reality captured in the image and as reality(ies) of the person(s) who created this photograph. Therefore, the so called “picking and choosing” (Lincoln et al., 2018) and switching between these two sets of underpinning beliefs within a study becomes highly problematic for visual data and is therefore not, methodologically speaking, a sound choice. This being the case, tourism, hospitality and leisure researchers using visual methodologies through a reliance on visual or multimodal data sets which include visual data, ideally need to have as clear as possible positioning within each of their projects in as far as the core questions of ontology, epistemology, methodology, methods and axiology are concerned.

In this sense, what is commonly seen within visual tourism, hospitality and leisure research projects are different approaches in research design, methodology, methods and data sets preferred by researchers. Visual research projects in tourism, hospitality and leisure which are underpinned by positivism, will therefore incorporate hypotheses and large data sets (big data) in their attempt to produce objective, generalisable knowledge which is perceived as (absolutely) true. In this context visual data included in such projects will not only be a part of a large set of data, but each of the visuals included will also be perceived by the researchers conducting the project as reality which has been captured in these. Researchers whose projects are underpinned by post-positivism will still be perceiving visual data included in their studies as representations of a single reality which has been captured in that image, and there will be commensurability in that respect between positivism and post-positivism. That said, in the case of studies underpinned by post-positivism it is more likely to see an addition of qualitative methods to the largely quantitative research designs, and an assumption that knowledge created as a result of these projects is (probably) true and as such applicable to many contexts, and not only the specific context of the study in question.

Simultaneously, the group of paradigms seen as commensurable between each other but otherwise incommensurable with the positivist and post-positivist positions, are those paradigms which are broadly speaking informed by constructivist/constructionist worldviews and which among them include critical theory, constructivism and participatory approaches (see Table 1 below). Researchers working within these philosophical positions will have an entirely different view of visual data within their projects. Visual data in such projects will be treated as representation(s) of reality(ies) as perceived by the person(s) who either created, or in the case of elicitation techniques, experienced these. Importantly, the visual data will also often be contextualised within the wider context(s) of time, space and (inter)subjectivities, all of which will assist researchers in achieving the wider aims of their studies as guided by their respective axiological positions. In that regard, a reliance on predominantly qualitative research methods across the different paradigmatic positions in this wider group of paradigms is likely to be seen as well as a reliance on smaller sets of data (small data), and findings which, even if these may have implications for other similar contexts, are nonetheless seen as solely applicable to the context of each of these studies.

Therefore, whether authors have explicitly stated their philosophical positions or not, a perceptive reading of each study combined with a reflection on the table above is likely to provide sufficient insight to the reader as to which of the key paradigms or philosophical positions underpinned these. Such reading will also provide illuminating insights into the reasons why small data and big data approaches are rarely combined in the same project. An exception to this can usually only be found in projects conducted by researchers who hold a pragmatic position in their wider research work and thus find themselves switching from one paradigm to another in different studies (for a recent example of such pragmatism within a collection of projects aiming to provide insights into the same topic from different perspectives see Huang, 2025).

Data sets of visual research projects in tourism, hospitality and leisure studies are often multimodal, in that the visual is more commonly included and analysed alongside textual and/or numerical data rather than on its own (Rakić and Chambers, 2012b). Visual data can also be collected, introduced, created (Rakić and Chambers, 2012b) as well as analysed (e.g. see Rose, 2023) through a reliance on a multitude of different methods, which are more commonly, but not exclusively qualitative. Until relatively recently, academic knowledge was predominantly marked by studies designed to collect data which would answer specific research questions oriented towards deep understanding – or, in other words, small data approaches (Kitchin and Lauriault, 2015). Facilitated by growth in both access to massive volumes of online data, and availability of advanced technology for analysis [e.g. artificial intelligence (AI) such as machine learning and deep learning], the visual turn is advancing in new directions, with the rising popularity of big data approaches. Namely, in contrast to small data, big data sets are marked by their very large size, and more often than not an exhaustive sample which contains secondary data collected online either in real, or near-to-real time, for the purposes of analysis (Kitchin and Lauriault, 2015; Kitchin, 2013).

Notwithstanding the importance of philosophical underpinning of both small and big data approaches in research, the continuously expanding opportunities that social media create as platforms from which visual data can be collected has been noted by a number of scholars especially given their potential to provide a novel understanding of social phenomena (Volo and Irimias, 2021). In the context of tourism this can include, for example, advancing our understanding of tourists’ preferences, interpretations and behaviour (Deng and Li, 2018; Höckert et al., 2018) based on analysis of visual data collected from a variety of different social media platforms. Some of the most commonly relied on sources of (big) visual data include platforms such as Flickr, Twitter, Facebook, Instagram, TikTok, other publicly available tourism related platforms and resources such as Airbnb, TripAdvisor, Yelp and destination and tourism businesses websites that incorporate user-generated content (Xiang and Pan, 2017). Although visual data is wide in scope and can include photographs, videos, drawings, artwork, memes and other types of visual representations, the most popular (big) visual data type tends to be still image in the form of photographs (e.g. see Cohen et al., 2022; Pantano and Dennis, 2019; Yu et al., 2020; Zhang et al., 2024).

Within tourism research many initial contributions used mostly meta data such as user profile, user network, reputation status (e.g. badges in social media platforms), date, time and location of the visuals and did not truly exploit the content of the visual itself (e.g. Bui et al., 2022; Zheng et al., 2023). Similarly, within hospitality research there is an overreliance on photographs as data types in big data studies. That said, the hospitality sector is characterised by extensive user-generated reviews available online and guests share photographs, and at times also videos, alongside textual reviews expressing their evaluation of the experience of lodging and restaurants. This being the case, there is certainly potential to expand the visual data types collected in future big data studies in both tourism and hospitality studies.

Tourism scholars frequently address topics such as destination image (e.g. Xiao et al., 2022), posting and sharing behaviour (e.g. Cai et al., 2024), decision-making (e.g. Zheng et al., 2023), evaluation and rating (e.g. Liu et al., 2024), engagement (e.g. Mariani, Mura and Di Felice, 2018), attractiveness (e.g. Pantano and Dennis, 2019) and emotional perception (e.g. Zhang et al., 2024). The latest advancements in data analytics have offered opportunities to analyse complex data on a large scale (e.g. up to hundreds of thousands of visual images) that feeds computational models. Data collection is mostly done through data scraping (e.g. self-developed software using Python) (Liu et al., 2024). Typically, big data collection and analysis involve modules such as extracting data, data processing and aggregation and visualisation (e.g. Mariani et al., 2016). More specifically, analytical methods used include image analysis with extraction of features such as colourfulness, or the identification of objects or human faces (e.g. using Google Cloud Vision API). Data is then used for regression analysis (Liu et al., 2024; Mariani et al., 2016) and econometric modelling (Cai et al., 2024) or fed to machine learning methods and algorithms for visual content and sentiment analysis (Bui et al., 2022; Zhang et al., 2024) or explored using cluster analysis and plotting images on a map (Pantano and Dennis, 2019) or with multilayer neural networks (Zheng et al., 2023), and scene recognition and semantic classification (Zhou et al., 2017).

In hospitality studies, on the other hand, scholars often call for big data studies that would shed further light onto a deeper understanding of the impact that visuals have in a variety of hospitality specific contexts (Kwok et al., 2017), such as hotel and restaurant reviews. Common themes of research revolve around effects that photographs, attached to reviews, have on consumers/tourists trust and purchase intention (Liu et al., 2022; Park et al., 2021; Zinko et al., 2020), enjoyment (e.g. Yang et al., 2017) and perceived usefulness of the content (e.g. Li et al., 2023; Ma et al., 2020; An et al., 2020). Among a variety of related themes, aesthetics of photographs has also been explored in relation to the impact on pricing strategies of Airbnb (Hu et al., 2023), as were the effects of user generated photographs on pricing for hotels (An and Ozturk, 2022) alongside the effect of visuals on consumer engagement (Hou and Pan, 2023). Interestingly, and linking these debates to the philosophy of science and how visual data is perceived when viewed from different paradigmatic lenses or worldviews (see Table 1), some hospitality scholars have explored the effect of misleading photographs which were published on hotel websites, showing how anger, mistrust and regret may result from a reliance on visuals to create false beliefs (Kuo et al., 2015; Khan et al., 2024). Other hospitality studies have also included a focus on facial expressions and emotional clues within photographs to evaluate perceived benefits with respect to peer-to-peer accommodations and possible mitigation of risk perceived when booking (Ert et al., 2016). Such studies will often include object detection techniques to synthetise the content of the photos and then regressions or econometric analysis and/or deep learning to exploit the value of visuals.

Within the wider visual turn, big data approaches within visual research projects in tourism, hospitality and leisure studies are more reflective of positivist and postpositivist philosophical underpinnings (see Table 1). Although, big data approaches are more common in tourism and hospitality studies than in the field of leisure studies where small data studies, many of which are qualitative, continue to dominate the research landscape (e.g. see special issues Pernecky and Rakić, 2019; Todd et al., 2025). This point is particularly significant in that although big data approaches with their inherent positivist and postpositivist philosophical positions are on the rise within the wider visual turn, small data approaches are by no means being replaced by big data. In fact, small data approaches underpinned by a wide variety of philosophical positions and methodological approaches (see Table 1) are expected to continue flourishing both across disciplines and fields of studies (Kitchin and Lauriault, 2015) as well as within the specific context of visual research methods (e.g. see Rose, 2023), while being complemented by (in as far as knowledge creation is concerned) rather than replaced by big data approaches.

Vibrant advances in small data approaches are also growing, studies that use unstructured arts-based approaches such as drawings to shed light on how people perceive places (Matteucci and Önder, 2018), offering new views on imaginative and affective associations with places (Köstinger and Matteucci, 2024), represent examples of how novel designs to small data approaches can help overcome challenges from physical, socio-cultural or linguistic distance in data collection. Further to this, academic filmmaking and video production in these fields of studies is also on the rise (e.g. the recently produced nature matters by Relton (2024) and the dark matters by Burns (2024) in the context of leisure studies and the embodied absence of the past by Akwasi Adu-Ampong (2025) in the context of heritage and tourism studies) as are embodied methodologies, audio-visual impressionistic tales and netnographic snippets which have been used to explore train travel phenomena (Jensen et al., 2015). Similarly, interpretive and participatory netnography was used by Conti and Lexhagen (2020), to extend knowledge on the role of online photography in creating experience value in nature-based tourism.

Numerous studies in the fields of tourism, hospitality and leisure continue to rely on a wide variety of small data approaches (e.g. see relevant contributions in Carnicelli et al., 2017), some of which increasingly incorporate both offline and online data created for the purpose of these projects and with specific research questions in mind. Namely, small data studies can rely on qualitative, quantitative or mixed methods approaches. Typical of the qualitative small data approaches in the context of leisure studies for example, methods such as photo-elicitation techniques have been incorporated within an ethnographic study of a farmer’s market in an urban environment (Johnson, 2014). In another example, Q method with photographs and photovoice have been used to study leisure practices of older individuals (Annear et al., 2014) and, more recently, a freehand drawing method has been used within focus groups within a research project focusing on serious leisure (Wyatt et al., 2025).

While aiming to provide some key insights into possible futures for the visual turn in tourism, hospitality and leisure, it is important to highlight that these concluding remarks are not meant to be neither prescriptive nor favouring particular approaches. In other words, we acknowledge important contributions to wider knowledge production by studies underpinned by both core groups of paradigms or worldviews, both the “positivist” and the “constructivist/constructionist” worldviews, as well as by both small and big data studies. It is in fact this very richness and variety of different approaches which is expected to continue fuelling the visual turn in tourism, hospitality and leisure studies for some time to come.

Specifically, data visual studies in tourism, hospitality and leisure are likely to continue rising in their popularity over time, with debates over the role and ethics of AI in the various stages of the research process – ranging from data collection and analysis to visual research outputs relied on to represent the findings. The phenomenon of scholars in these fields of studies engaging with both the visual studies publications coming from their core discipline(s) and equivalent literature coming from their subject specific multidisciplinary field of studies, will add further richness and variety of different ways in which both big and small data studies are designed in the future. In that sense, we are likely to see more creativity and less prescriptiveness in as far as options available to visual tourism, hospitality and leisure researchers within methodology focused publications.

Simultaneously, we are likely to see a maintained space for knowledge production of small data studies, only some of which will completely exclude visual data sourced from online sources, given that increasingly lives are lived both in the real and the virtual world. Therefore, what we are likely to see in the context of small data studies in these fields of studies is increasing incorporation of data sourced from online sources, with some small data studies such as digital ethnographies/netnographies conducted entirely online. In that sense, we are likely to see increased creativity across the big/small data divide with researchers increasingly experimenting with research outputs such as exhibitions, film and video productions and AI generated images which are relied on to represent research findings. With respect to the collection and analysis of both digital and non-digital or material visual data, a further avenue for future research entails the ongoing debates on the ethical risks associated. Overall, the possibilities of visual research are endless, and it is up to scholars in tourism, hospitality and leisure studies to embrace these possibilities across the big data/small data divides and, of course, within the different paradigms underpinning their research.

Serena Volo acknowledges the Free University of Bozen research fund: Climate crisis in Alpine tourism: A mixed-method investigation through Visual Social Media, I53C22001190005.