Keywords

1 Introduction

Tourism is a significant economic factor for many regions and accounts for a considerable amount of movement within and between regions. Therefore, tourist flow, which refers to the projection of the trajectory of tourists and their activities in geographical space [1], is of great importance for the tourism industry and has become a highly researched area. The development of location-acquisition technologies and the rising popularity of Online Social Networks (OSN) led to the availability of a huge amount of data that provide very useful information on spatiotemporal behaviors of individuals as well as corresponding interactions between users. When travelling, tourists nowadays typically use their smartphones to upload photos and videos with text information to numerous social media channels, documenting their trip in real-time. The use of location-based social network (LBSN) data can help us to recognize potential cases of overcrowding and other problematic trends at an early stage to avoid conflicts that could stem from the tourist flows before the problem areas fully manifest. However, prior research of digital footprints has not included social network interactions and therefore ignored awareness and popularity of posts. This paper assumes that these interactions are proxys of interest in a region and therefore might be an indicator of the intensity of tourism flows.

2 Literature Review

Today, the majority of tourism processes and transactions are digital. Most travelers use online tools and social media to plan their stay, to acquire information before and during the trip, to book a hotel or transportation, and finally to share their experiences with others. Therefore, the importance of OSNs has increased significantly in the last decade [2] and became one of the main sources used by tourists for compiling and sharing information. Furthermore, the rapid growth of mobile devices such as smartphones and tablets and the increasing availability of location-acquisition technologies including GPS, Wi-Fi and 5G, allow users to publish media content along with their position as location-tagged media content. In this way, social networks become Geosocial or LBSNs [3] enabling researchers and marketers to obtain location-based user information. LBSNs include Facebook, Flickr, Foursquare, Google+, Instagram, and Twitter. The identification of the spatiotemporal distribution of tourism flows and the recognition of popular touristic sites from LBSN data is one of the most researched topics in tourism. For instance, Peng and Huang [4], Su et al. [5], and Zhou et al. [6] proposed methods to find tourist hotspots while Mou et al. [7], Vu et al. [8], and Wu et al. [9] showed how tourism flows could be extracted from geo-tagged Flickr photos.

Another stream of research combines spatial and semantic data to analyze tourists’ preferences. Brandt et al. [10] demonstrated that SMA captures spatial patterns within the city that reveals environmental and topical engagement. Miah et al. [11] combined four techniques (text processing, geographical data clustering, visual content processing, and time series modeling) on Flickr data to analyze tourist interests, trends, and seasonal patterns. Shi et al. [12] exploited tourism crowding from crowdsourcing geospatial data, and Jiang et al. [13] investigated the tourist sentiment changes between different attractions based on geotagged social media data derived from Sina Weibo.

However, among the many studies analyzing LBSN data, there are only a few linking geotag metadata with other information. To achieve a more complete picture of travel patterns and location attractiveness, this paper proposes a new methodology linking geotagged pictures with online reactions from users. First, geotagged pictures of single users provide insight into their travel behavior. Second, the resulting travel paths are enriched with comments and likes to act as a proxy for the users’ awareness of joint locations emanating from a LBSN platform like Instagram.

3 Case Study Background and Data Collection

The most western province of Austria, Vorarlberg, includes six defined destinations, of which Montafon valley is one of the top three in terms of arrivals. The DMO is promoting their hashtag #MeinMontafon (45.8K posts) on their own Instagram account, which has 20.8K followers as of July 2022. 18,504 public Instagram posts containing the hashtag #Montafon were acquired from Picodash (https://www.picodash.com) in April 2022 (2022: 5,428; 2021: 13,041). The following variables were processed: userID, geotag information, and # of comments/likes.

4 Methodology

In a first step, 7,521 geotagged posts assigned to a location within the case study border were kept (see https://touren.montafon.at/en/tours/). As users could assign multiple posts to the same location, the number of comments/likes collected by each user (userID) for each specific location were aggregated, 4,295 location-userID combinations. Instagram users who only linked posts with one single location were deleted, 2,460 posts.

In the second step, two different weighting procedures were applied to the remaining posts. The first one, called path dominance, is based on pairs of locations with posts by the same user. Occurrence frequencies for each pair were determined through all users. The second/third weighting, social media presence, is determined by the sum of comments/likes collected through all users attributed to each location pair determined before.

5 Results

Following visualizations, using the R package ggmap [14], illustrate the 14 pairs with the highest number of occurrences and comments/likes respectively. Path thicknesses of Fig. 1 come with a minimum occurrence of 20 or higher (grey coloured in Table 1). One path showed up 30 times, two paths 26 times, etc.

Table 1. Path occurrence–Montafon.
Full size table
Fig. 1.
figure 1

Path dominance–Montafon.

Full size image

Figure 2 displays each pair’s attention attracted on Instagram. The maximum number of comments/likes for a path was 1,915/38,718. Path thicknesses of Fig. 2 come with a minimum of 597/17,215 comments/likes for each location pair.

Fig. 2.
figure 2

Social media presence based on comments (left) and likes (right)

Full size image

6 Discussion

LBSN posts reveal insight into travel paths of tourists. Depending on the means of investigation, either raw path trajectories, or measures including social media attention (comments and likes) can be observed. Concordance between a DMO’s marketing strategy and the representation of a destination’s locations on social media platforms is a must to pool both advertisement mediums.

The DMO of the case study region Montafon has formulated their brand slogan as “Real mountains. Real experiences”. In terms of local attractions motivating tourists to visit the destination, the DMO has, in line with their slogan, put a specific emphasis on the mountains surrounding the valley and its villages [15]. Consistent with this, the most common path (Figs. 1 and 2) includes the mountain range dividing the two villages with the highest altitude Gargellen and Gaschurn. Observing Fig. 1 by post frequency, the path lies between the two mentioned villages. The strongest path in correspondence to comment engagement (Fig. 2) connects afore mentioned mountain range and the mountain Kristberg/Innerberg in the village Silbertal. The second most frequent path is again the mountain range between Gargellen and Gaschurn with its village Bartholomäberg. The latter path is simultaneously the most common path by likes, together with a path to the high mountain lake Lünersee in the village Vandans. In general, the distribution of paths varies between frequency, comments and likes. However, it can be deduced that travel paths determined from the LBSN Instagram are in line with the DMOs’ advertising messages.

Summarized, different path determination strategies offer a more comprehensive picture of customer perceptions communicated on social media channels. Comparison of this new approach with traditional tourism flow analyses will be necessary to assess the relevance of the inclusion of social network interactions.