1. Introduction
The movement of tourists is a key factor in tourism and also represents very important information in terms of understanding tourist behavior and the role of specific tourist destinations [
1,
2,
3]. Specifically, because the number of individual tourists who do not rely on travel agencies and make their own travel decisions instead is increasing, tourist movement patterns are becoming more and more complicated. The development of the Internet and information and communication technology has made tourists smart and independent. Mobile devices have created an environment in which tourists can create and consume a great deal of information, as well as sharing information in real time [
4]. Therefore, the guidebook and tour guide are being replaced by mobile devices. These new information-sharing activities are going on constantly, without any restrictions based on the physical environment. In this context, this study examines such individuals’ independent travel behavior in terms of sustainable tourism. Sustainable tourism has mostly been studied in terms of development in an attempt to determine how tourism destinations should be developed in harmony with local residents and the natural environment [
5,
6,
7,
8]. However, focusing on tourists can provide infinite information without environmental destruction and the accompanying tourism activities can also be viewed in a broader sense. Travelers who enjoy traveling independently using information obtained via the virtual space information are more concerned with maintaining harmony with local residents as compared to group tourists and seek local authenticity [
9].
As mentioned above, the development of the Internet has strengthened the information for search among tourists. Because the number of individual tourists who do not rely on travel agencies and make their own decisions instead is increasing, tourist movement patterns are becoming more and more complicated.
Given this trend, tourists who spend more time at a destination find it worthwhile because it allows them time to seek new experiences, choose destinations based on their cultures and visit more tourist destinations in total [
10]. Furthermore, platforms to help tourists who plan their trips alone are emerging. For instance, Fortune Korea 2014 introduced Stubby Planner, an innovative tool. Now, one million people use this platform to plan their trips to Europe annually.
According to Cohen [
9], these tourists can be termed ‘noninstitutionalized tourists.’ They take trips freely and visit multiple tourist destinations. A typical example of a tourist with these characteristic is a backpacker. A backpacker is a young tourist who travels without the advice of a tour guide and is free to plan his or her own schedule [
11,
12]. Because of this freedom of movement, backpackers generate more complex movement patterns than tourists in large groups, such as mass tourists, who are generally moving on the same itinerary and these movement patterns may contain more useful information for tourism marketers and developers. Many researchers have studied such tourists’ movement patterns and suggested various implications [
12,
13,
14,
15,
16]. However, despite the increasing the number of backpackers in Korea and the popularity of backpacking in Europe, there is still a lack of understanding of their movement patterns. Therefore, the present paper aims to investigate the movement patterns of Korean backpackers in Europe by using Social network analysis. First, we measure the movement patterns of Korean backpackers in Europe in 2015. In doing so, we attempt to identify key tourist cities in Europe from the perspective of Korean tourist and understand the connections between tourist cities. Second, we attempt to analyze the changes in Korean tourists’ movement patterns by comparing the 2015 data with the 2012 data. Finally, we suggest ways of developing of more efficient and productive tourist infrastructure.
2. Literature Review
2.1. Tourist Destination
Tourist destinations are a crucial factor in the tourism industry [
17]. However, the concept of a tourist destination is vague and broad. For instance, an attraction, such as Disneyland, can be a tourist destination and a city, such as Rome, can also be a tourist destination. Therefore, it is important to clarify the definition of a ‘tourist destination’ in order to understand tourist movement patterns.
Leiper [
2] discusses tourist destinations as one of the geographical elements of tourism. Leiper [
2] defines a tourist destination as a location that can attract tourists to visit it. Thus, his definition of a tourist destination is well-characterized but also abstract. Lew and McKercher [
18] argue that tourist destinations involve various factors, which can be divided into primary attributes and secondary attributes. Primary attributes are characteristics that are inherent to a tourist destination, such as its ecology and culture. Secondary attributes are characteristics that are created via development, such as hotels. These attributes combine to make a given tourist destination attractive to tourist destination [
17]. However, the geographical discussions of tourist destinations remain vague.
The WTO [
19] attempted to create a concrete definition of a tourist destination. According to the WTO [
19], a local tourist destination can be defined as “a physical space that includes tourism products, such as support services and attractions and tourism resources. It has physical and administrative boundaries defining its management and images and perceptions defining its market competitiveness. Local destinations incorporate various stakeholders, often including a host community and can nest and network to form larger destinations. They are the focal point in the delivery of tourism products and the implementation of tourism policy ([
18], p. 405)”. The purpose of the present study is to measure tourist movement patterns between various destinations in Europe. Additionally, because Europe is largely urban, many cities are represented among tourist destinations. Thus, the present study defines a tourist destination as a city, including the products and activities in that city that attract tourists.
2.2. Tourist Movement Patterns
From a geographical point of view, tourists who visit more than one tourist destination create spatial movement patterns. These tourist patterns can be global, national, or local [
20] and can include a variety of information that is useful for tourism marketers [
1,
3,
14]. A concrete understanding of the spatial movements of tourist can provide insights into tourist behavior, including the destination characteristics that are most attractive to tourists [
21]. Lue, Crompton and Fesenmaire [
11] examined the effect of the spatial patterns of attraction on tourist routes and conceptualized multi-destination trip behavior. Other researchers have focused on tourist movement patterns rather than the locations of tourist destinations. Specifically, Pearce [
12] suggested that the direction of transportation development should be based on the analysis of tourist movement patterns in Europe. Shin [
22] examined the movement of automotive tourists in Taiwan and found that they engaged in various movement patterns and that 16 tourist destinations all had different roles. Leung et al. [
14] investigated the movement patterns of international tourists who visited Beijing, including the effect of 2008 Beijing Olympics on tourist movement patterns. According to his study, these tourists tended to visit famous traditional attractions and their movement patterns were concentrated in the central city area. In addition, tourists who visited during or after the 2008 Beijing Olympics showed extended movement patterns as compared to tourists who visited before the Olympics. On the other hand, the inherent characteristics of tourists can also influence their movement patterns. Lew and MaKercher [
18] argue that tourists’ movement patterns reflect their consumption styles. Hwang et al. [
13] also explored international tourists’ travel patterns within cites in the United States. According to his study, these patterns differ with tourists’ origins and levels of familiarity with the US. Specifically, Asian tourists tended to visit Los Angeles, Las Vegas and San Francisco, whereas European tourists and New Yorker tended to visit Orlando, Miami and New York, which had low levels of popularity among Asian tourists.
Cohen [
9] distinguished between two types of tourist in terms of sociology. First, institutionalized tourists depend on a travel agency or guide. They enjoy passive tourism activities and moving about as ordered. On the other hand, noninstitutionalized tourists attempt to experience the culture of any tourist destinations they visit. They enjoy active tourism activities and novelty. Backpackers are an example of noninstitutionalized tourists: “backpackers (are) predominantly young travelers on extended holidays with a preference for budget accommodation, a flexible and informal travel itinerary and an emphasis on meeting people and participating in a range of activities ([
23,
24], p. 194)”. That is, backpackers are young tourists who do not move according to the travel agency’s schedule and are free to plan their own schedules. They have a strong motivation to escape from their daily lives [
25], seek out unique sites and interact with the culture of tourist destinations they visit [
9,
26]. Therefore, backpackers can create their own routes and travel to more complex and remote tourist destinations than mass tourists. The development of information and communication technology has enabled tourists to share information, [
4] and thus, a great deal of tourism information has been generated. As a result, tourists are becoming smarter and more able to travel independently by using the real-time personalized information provided at tourist destinations. In this context, the present study examines the movement patterns of backpackers.
2.3. Korean Backpackers
Pizam and Sussmann [
16] explored tourist behaviors by nationality and found that Korean tourists preferred familiar places, rather than experiencing other cultures during overseas trips. Recently, however, several researchers have identified the characteristics of Korean tourists, especially backpackers in Korea, that are different characteristic from those mentioned above. Specifically, Park and Santos [
27] examined memorable tourism experiences among Korean backpackers in Europe through interviews and found that they chose European backpacking trips to experience different cultures. These tourists reported that a unique experience was most important in terms of having a memorable trip. Bae and Chick [
25] explored the characteristics of domestic Korean backpackers and found that Korea backpackers were mostly young people who wanted to take long trips during their school break and attempt to experience events that they could not experience in daily life. The number of young tourists has gradually increased [
28] and this is reflected in the analysis of travel agency product sales. According to GTN [
29], the sales volume for individual products, such as airline tickets and hotel products, is increasing rapidly as compared to package product sales. Among the seven biggest travel agencies in Korea, European airline tickets account for most of sales volume. Thus, one can infer that Korean backpackers prefer Europe as a destination for free travel.
3. Research Question
Tourist movement patterns, including spatial and temporal information, are key to understanding tourists [
1,
3,
14]. In particular, backpackers’ movement patterns are more informative for understanding tourists’ behavior than those of package tourists because backpackers visit more various destinations and take longer trips [
9,
24,
26]. The analysis of the spatial and temporal movement patterns of backpackers can help provide useful information about them. Consequently, this study aims to analyze the travel movement patterns of Korean backpackers in Europe.
Research Question 1. Which city was most central for Korean backpackers in Europe in 2015?
Research Question 2. Were there differences in city centrality for Korean backpackers in European networks between 2012 and 2015?
4. Method
4.1. Sample and Procedure
To analyze the movement patterns of Korean backpackers in Europe in 2015, this study collected secondary data. Specifically, this study was conducted using NAVER, one of the most popular portal and blog service sites, which has an approximate share of 71% in Korea [
30]. For the network analyses, postings on the blogs might be useful for comparison with photo-based SNS services because they provide the entire itineraries of each backpacker regarding European tourism sites. The researcher searched for blog postings using the term ‘Europe backpacking route’. Posts written during the one-year period from 1 January to 31 December 2015 were included. Of the 717 blog posts that appeared in the search, we excluded postings that contained advertisements or recommendations or did not mention detailed information. This left 122 postings that contained actual routes including specific cities. Thirty-four postings from 2012 were collected in the same way. This study also collected data from 2010 and tried to compare the trend of backpacking trips after five years. However, the number of blog posts in 2010 was too small and those posts provided little information about specific routes. Therefore, this study adopted data from both 2015 and 2012 (
Appendix A). From the 2015 data, 162 cities were identified and 129 cities were identified from the 2012 data. This study initially examined the top 20 cities Korean backpackers had the visited and found that Paris (France) was ranked first in 2015 and that London (UK) was ranked first in 2012 (
Table 1 and
Table 2).
4.2. Network Analysis
Tourists’ movement patterns can be examined through network analysis. Network analysis is a set of research procedures for identifying structures in systems based on the relationships among components [
31] Network analysis can be used to describe the global level of structure because it examines system indicators such as centrality, connectedness, integrativeness and system density, as well as the potential clustering of the network into subgroups.
The basic network dataset is an
n ×
n matrix S, where
n equals the number of nodes in the network. A node might be an individual or a higher-level component, such as an organization or a nation, of which the system is composed. Each cell
sij indicates the strength of the relationship between nodes
i and
j. In communication research, this relationship is generally defined by the frequency of communication between nodes [
32,
33,
34,
35]. To examine global corporate communications regarding the movement routes of tourists, this study used a 162 (cities) × 162 (cities) matrix for the 2015 network and a 129 (cities) × 129 (cities) matrix for the 2012 network. Each cell was weighted according to the frequency of moving from the city in the column to the city in the row. Therefore, nodes indicate European cities and links indicate the movements of Korean backpackers.
Next, using UCINET 6, degree centrality was calculated in order to determine which cities play significant roles in determining the movements of Korean backpackers in the European network. The key benefit of normalizing degree centrality in this way is that we can assess the relative centrality of two cities.
Degree centrality, in this study, indicates the number of co-visitors between two cities. Because centrality is a structural attribute of each node in the network, centrality identifies the central point based on many direct contacts with other points [
34,
36]. This study adopts the Freeman approach to calculating the degree centrality of each node, as well as the overall network degree centralization, because backpackers’ networks are significantly asymmetric. For non-symmetric data, the in-degree of a node
u is the number of ties received by
u, while the out-degree is the number of ties initiated by
u [
34,
37]. As defined above, degree centrality is simply the number of nodes that a given node is connected to. In this case, out-degree centrality is the number of backpackers that have left from a given city. In-degree centrality is the number of backpackers that have traveled to a given city. A higher number of visitors does not simply increase the degree centrality score. Rather, to obtain a higher degree centrality score, a city must be linked with certain other cities. In other words, the way in which cities build relationships with other core cities is crucial in positioning them at the center of the network of Korean backpacker routes in Europe.
Eigenvector centrality is an ideal measure for those networks in which the tie strength between actors, rather than simply the presence or absence of a tie, is known [
32,
33,
38]. It considers the strength of ties, including indirect social ties, among nodes. Thus, more central destinations can boost their centrality due to the inherent circularity involved in the calculation of the eigenvector centrality measure. This has the effect of making actors with strong ties to more central actors appear to be more central [
32,
33,
38,
39]. Betweenness centrality refers to the “share” of the shortest paths in a network that pass through a certain node [
40]. Thus, betweenness centrality can be affected by the number of cities that are specifically co-linked specifically with two other cities. Finally, in this study, a high closeness value indicates that travelers leaving a city require the minimum steps to reach all other nodes. The lowest possible score occurs when the node has ties to every other node. Therefore, if a city is the most central in terms of closeness, it should allow travelers to quickly reach all other cities in Europe.
Table 3 shows a summary of measurement methods mentioned above.
5. Results
The results reveal the structure of the network of Korean backpackers in Europe. Overall, Italian cities, such as Firenze, Venezia and Rome, are continually ranked in the top 20 in the list, regardless of year. Therefore, these cities play a key role for Korean backpackers in Europe. In addition, Paris (France), London (UK), Prague (the Czech Republic) and Interlaken (Switzerland) are also considered key cities. The specific results are described below.
5.1. Networks Structure of Korean Backpackers in Europe: Degree Centrality
Table 4 shows the centrality scores for the top 20 out of 162 cities in 2015. In this case, degree centrality indicates the total number of co-hyperlinked cities that one city shares with other cities. When the degree centrality analysis results for 2015 is compared with the results for 2012, differences emerge (
Table 4). Specifically, in 2015, Firenze (109, 2nd in 2012) had the highest out-degree centrality, followed by Venezia (107, 4th in 2012), London (100, 1st in 2012), Paris (94, 7th in 2012) and Interlaken (92, 9th in 2012). The in-degree centrality rankings are significantly different from the out-degree centrality rankings. In 2015, Rome (112, 2nd in 2012) had the highest in-degree centrality, followed Paris (110, 1st in 2012), Venezia (108, 4th in 2012), Firenze (107, 3rd in 2012) and Prague (94, 7th in 2012). The interesting result here is the difference between London’s in-degree and out-degree centrality. London’s the out-degree score was quite high, while the in-degree score was low in both 2012 and 2015 (
Table 5). Both the out-degree and in-degree centrality results show that Italian cities (i.e., Firenze, Venezia and Rome) rank at the top of the list in 2012 and 2015.
5.2. Networks Structure of Korean Backpackers in Europe: Eigenve Centrality
For 2015, the eigenvector centrality results were similar to those for degree centrality. Firenze had the highest eigenvector centrality (75.473), followed by Venezia (70.474), Rome (55.151), Interlaken (33.249) and Milano (31.685). However, the Italian cities (e.g., Milano and Pisa) moved up in the rankings as compared with Paris and London, which moved down in the rankings (
Table 6). Additionally, comparing 2012 and 2015, the eigenvector centrality results are similar and in 2012 (
Table 7), Firenze (1st in 2015) had the highest eigenvector score (67.063) in the rankings. However, there is a slight difference in the order in that in 2012, Firenze was followed by Rome (3rd in 2015), Venezia (2nd in 2015), Wien (7th in 2015) and Munich (13th in 2015).
5.3. Networks Structure of Korean Backpackers in Europe: Betweenness
Table 8 shows the results of the betweenness analysis for 2015. These results are somewhat different from those for the 2012 network (
Table 9). Specifically, in 2015, Paris had the highest betweenness score (28.702, 1st in 2012), followed by Rome (18.345, 4th in 2012), Venezia (14.113, 5th in 2012), Prague (13.573, 3rd in 2012) and Interlaken (12.699, 10th in 2012). Norwegian cities that had not been identified in 2012 entered the top 20 and Budapest (25th in 2012) moved up 14 steps in the rankings.
5.4. Networks Structure of Korean Backpackers in Europe: Closeness Centrality
High closeness centrality indicates that when leaving from a node, a traveler must take only the minimum number of steps to reach all other nodes. The results showed that Paris had the highest closeness centrality (29.87), followed by Munich (29.22), Interlaken (29.16), Prague (29.06) and Rome (28.95). These results are similar to those for betweenness in that four (e.g., Paris, Interlaken, Prague and Rome) of the top five cities are the same (
Table 10) and quite different from the results in 2012 (
Table 11).
6. Discussion
This study explored the movement patterns of Korean backpackers in Europe through network analysis. The main contribution of network analysis is to offer broad pictures of recent dynamic relationships, which are critical in terms of understanding tourist behavior among European tourism sites. Furthermore, this study attempted to identify cities that are characteristic of the 2015 network as compared with the 2012 network. Network analysis provides various methods with which to investigate and compare movement patterns. Overall, Italian cities played a key role in 2015 when Korean backpackers travelled to Europe as compared to 2012.
The out-degree centrality values in the 2015 network indicate that certain core cities—Firenze (Italy), Venezia (Italy), London (UK), Paris (France) and Interlaken (Switzerland)—played a key role in the movement patterns of Korean backpackers in Europe. However, the order was slightly different than that for in-degree centrality. In particular, London (UK) had a significantly lower score for in-degree centrality than for out-degree centrality. This means that more tourists move from London to other cities than form other cities to London. In other words, Korean backpackers tend to choose London (UK) as their first city when they travel to Europe. In contrast, Rome (Italy) had an in-degree score that was higher than its out-degree score. This means that Korean backpackers tend to choose Rome (Italy) as their final city. Furthermore, as compared to 2012, in 2015, Italian cities were ranked more highly. Thus, Italian cities are playing a key role for Korean backpackers traveling in Europe. Among flights from Korea to Europe, there are 14 flights to Paris and 13 flights to either Rome or Milan. On the other hand, there are four flights to London, where Korean tourists have chosen to start their European trips [
41]. Because the United Kingdom is geographically distant from other European countries in terms of location, backpackers may feel that travel there is relatively difficult.
The eigenvector centrality results for the 2015 network show that Firenze (Italy) had the highest score, followed by Venezia (Italy) and Rome (Italy). Unlike degree centrality, eigenvector centrality does not simply represent being connected to many other cities. Rather, it identifies cities that are connected to core cities. Firenze is linked to key cities such as Munich (Germany) and Salzburg (Austria). Therefore, Firenze (Italy) is the most influential city in that it is connected to the core cities. Moreover, the top three cities were all Italian cities (i.e., Firenze, Venezia and Rome). The results are the same for the 2012 network. In 2012, although the order is slightly different, the top three cities are all Italian (i.e., Firenze, Rome and Venezia). Thus, most Korean backpackers travelling to core cities tend to visit Italy and Italy is the most important tourist destination for Korean backpackers.
On the other hand, the betweenness results show that Paris (France) had the highest scores in both 2012 and 2015. Betweenness indicates the extent to which cities are not directly connected. For example, Paris (France) plays a role in connecting the overall cities in the network. That is, Korean backpackers are the most dependent on Paris (France) when traveling to other cities. In the case of Paris, the frequency of visits is high and the distance to London, which is the typical starting city for European travel, is also short. In addition, unlike UK cities that require air travel, Paris can be reached via a rail option called Eurostar. Eurostar’s London-Paris section has the highest number of sales and the top ranking in Korea [
42]. Thus, Korean backpackers who have flown into Europe through London can be found traveling through Paris to other European cities. On the other hand, as compared to the 2012 network, in 2015, the role of Italian cities became more important. Specifically, Rome (4th, Italy) and Venezia (5th, Italy) moved up in the rankings. Paris (France) plays a central role for Korean backpackers in Europe and serves as a link between cities that are not directly connected.
7. Conclusions
This study investigated the routes of backpackers in Europe. In particular, it examined specific movement patterns between European cities. Tourists’ movement patterns are very important in understanding tourists because they contain a great deal of information [
1,
2,
3]. This study identified the cities (e.g., Venezia, Paris, London, etc.) that are preferred by Korean backpackers, as well as the cities (e.g., Paris) that play a major role in tourists’ movement.
This study also has practical implications. First, this study identified the key cities for Korean backpackers in Europe. More specifically, Korean backpackers have traveled to Europe, mainly via London and have confirmed that they typically enter Continental Europe via Paris. Because backpackers in Korea likely prefer convenient transportation within Europe, destination marketers should design marketing that emphasizes convenient transportation in Europe. Second, a number of backpackers start from London, but London flights are fewer in number than those to Rome and Paris. Thus, it is suggested that Korean airline managers may be able to increase airline revenue by increasing the number of flights. In addition, Korean backpackers have been shown to move through Paris when traveling from London to continental Europe. This seems to be influenced by Eurostar and travel agency managers should be able to draw the attention of backpackers who do not typically rely on agencies by concentrating on those who purchase London-Paris Eurostar tickets. Lastly, as the era of smart tourism evolves, tourists’ movement patterns are becoming salient information. More specifically, as tourists’ accessibility to information increases, a variety of start-up companies are emerging that provide information on travel routes for tourists. It is suggested that, for businesses, this data on travel patterns can be useful.
Notably, this study has certain limitations, including that it inferred the factors that affect tourists’ movement patterns yet did not verify them empirically. Therefore, future studies can provide richer implications if they address the factors that influence travel routes and empirically verify the relationships between them. Moreover, this study collected information via specific blogs (i.e., NAVER blogs). Although NAVER has the highest share of Korean blogs, using their data cannot be generalized. Therefore, future research needs to collect information through multiple channels.