287Nod, G. and Aubert, A. Hungarian Geographical Bulletin 71 (2022) (3) 287–299.DOI: 10.15201/hungeobull.71.3.5 Hungarian Geographical Bulletin 71 2022 (3) 287–299. Introduction The tourism industry has a long-standing de- sire to predict the movement of tourists within and across destinations, their consumption and the popularity of tourist destinations. Researches focusing on this have used a va- riety of approaches. There are many studies on consumption theory, where the aim is to assess tourists’ choices based on their consump- tion decisions (Bernecker, P. 1962; Kotler, P. 1967; Mas-Colell, A. et al. 1995; Csapó, J. and M. Császár, Zs. 2021; Telbisz, T. et al. 2022) and their spatial movements (D’Agata, R. et al. 2013; Pécsek, B. 2015; Asero, V. et al. 2016). Also, the role of consumer preferences in travel decisions has been studied for decades (e.g., Woodside and Lysonski‘s “General destination choice model”) (Woodside, A. and Lysonski, S. 1989). The former use economic and socio- logical perspectives, but the researches also in- clude geographical approaches, for example, the optimal road accessibility of a given tourist destination (Tóth, G. and Dávid, L. 2009). In recent years, network approach research has become popular (Scott, N. et al. 2008a; 2009; Madarász, E. and Papp, Zs. 2013; Casanueva, C. et al. 2014), and thanks to digital advances we can now work with big data and determine the location and movement of an individual based on GPS coordinates or cell phone cellular data (Spinney, J.E. 2003; Ahas, R. and Mark, U. 2005; Díez-Díaz, F. et al. 2007; Wind, S. 2015; Zheng, W. et al. 2017). We have therefore seen several attempts to measure the number of visitors. At the moment, Hungary is in the process of in- troducing the registration and mandatory data reporting in the system of the National Tourism Data Supply Centre (NTAK), as a supplement to the “Government Decree 239/2009 (X. 20.) on the detailed conditions for the provision of accommodation services and the procedure for issuing accommoda- tion operating licences”. The obligation to provide data mainly concerned accommo- 1 Department of Tourism, Institute of Geography and Earth Sciences, University of Pécs. Ifjúság u. 6. H-7624, Pécs, Hungary. E-mails: gabica13@gamma.ttk.pte.hu, aubert@gamma.ttk.pte.hu Methods for measuring the spatial mobility of tourists using a network theory approach Gabriella NOD1 and Antal AUBERT1 Abstract The present study uses the methodological tools of network theory to investigate the spatial movements of tourists in the sample area, which is the South Transdanubian tourism region of Hungary. The basic idea of the study is that tourist movements across settlements in a larger tourist destination make a coherent network. As long as the approach is correct, this network can be measured by properties that are characteristic of net- works, such as centrality or degree. A review of the methodology of similar studies previously published on the subject has been used to supplement the method of analysis used below. As a result, the study not only characterised the sample area municipalities in terms of network characteristics, but also classified them into clusters for strategic planning purposes on the basis of the mobility propensity of the tourists staying there. Keywords: destination management, tourism mobility, cluster analysis, network theory Received March 2022, accepted August 2022. Nod, G. and Aubert, A. Hungarian Geographical Bulletin 71 (2022) (3) 287–299.288 dation establishments and is now gradually being introduced for catering establishments and operators of tourist attractions as well. In the long term, the obligation to provide data now in the implementation phase will provide a database for tracking the consumer choices of guests, and will also form the basis of the process discussed below, which could also be a tool for promoting the development of a destination marketing strategy or a de- velopment plan at regional level. The litera- ture on tourism destination management and tourist mobility; in Hungary (Piskóti, I. 2007; Sziva, I. 2014; Nod, G. et al. 2019; Aubert, A. et al. 2021) is well established, but the sys- tem is just in a phase of transformation, in Butler’s life cycle model (Butler, R.W. 1980), it is in the process of repositioning itself in the National Tourism Development Strategy 2030 – Tourism 2.0 (2021), which designates 11 tourist regions (Government Decree 429/2020 [IX. 14.]) within the borders of the country (Hungarian Tourism Agency 2021). The method presented in the study supports the development of regional strategies and the optimisation of the management system. The following study applies the network approach, one of the many theories outlined above, to investigate the mobility of tourists arriving in the South Transdanubian tourism region of Hungary during their stay in the re- gion, with the aim of determining the visitation of certain destinations, the length of stay in the region and the degree of mobility of tourists during their stay and in which direction. The theory is not alien to the literature. Asero, V. et al. (2016) conducted a similar study in Sicily (with a similar sample to the one used in the present study), with the aim of identifying tourism networks by analys- ing tourism mobility across destinations. The results show that tourists’ choices define the role of a destination within a network as “central” or “peripheral”. Similar clusters, but delineated on the basis of several vari- ables, are formed by the findings discussed later in this paper. A fundamental difference between the results of the two studies is that while Asero, V. et al. (2016) use a functional approach, the present study uses a more geo- graphic approach and also characterises the municipalities under study by other network characteristics based on tourism mobility. Since network research not only allows the discovery of network patterns, but also the prediction of their future functioning by knowing the properties and behaviour of the network (Scott, N. et al. 2008b, 2009), it is an excellent tool for strategic planning. The literature on network theory can be traced back to the Swiss mathematician Leonard Euler (1741), who interpreted graph theory in the urban spatial structure by using the logical pattern of the bridges of the former Prussian town Königsberg (today Kaliningrad, Russia) and road network. The tourist move- ments examined in this study also connect several settlements, i.e., they are measurable physical movements. However, they do not fol- low optimal public, rail or water routes, but ‘ar- tificial’ routes formed by tourists’ motivations. The discovery of Euler, L. (1741) laid the foundations for the development of graph theory, for which the terminology and formal tools were created by the Hungarian math- ematician Dénes Kőnig (1936). The random- ness and complexity of the networks that surround us in reality were described by the Erdős-Rényi model (Erdős, P. and Rényi, A. 1960), followed by the Watts-Strogatz model, which detected the formation of groups within the network (Watts, D.J. and Strogatz, S.H. 1998). The most recent major success in the study of networks was achieved by the re- search team led by Albert-László Barabási (2002), who described scale-independence using the example of the World Wide Web (Albert, R. and Barabási, A.-L. 2002). According to the hypothesis formulated during the research design, the role of each destination in the supply market of the region can be measured by the tourist movements of tourists across destinations (H1). To measure this, the study uses the centrality and degree calcula- tions of network theory developed by Albert, R. and Barabási, A.-L. (2002), Letenyei, L. (2006), and Barabási, A.-L. (2016) to deter- mine the centrality of the settlements. 289Nod, G. and Aubert, A. Hungarian Geographical Bulletin 71 (2022) (3) 287–299. The research designed to support this hy- pothesis is based on the following three re- search questions: Q1: Can tourism movements in the region be understood as a coherent network? Q2: Can the characteristics of the network (centrality, number of degrees) be used to characterise the position of a destination within the region? Q3: Does the mobility of the visitor staying in a given settlement characterise the market role of the settlement? Two studies are presented in the paper to support the hypothesis and answer the ques- tions. The first presents the extent of the net- work of connections of each destination (with other settlements in the region) using central and degree measures. The second part of the research characterises and classifies each set- tlement into separate clusters based on the mo- bility willingness of the guests staying there. Readers interested in the study but less familiar with the subject are recommended to consider the list of basic terms used later: Centrality – the most obvious measure of centrality is the number of connections (degrees) of each point relative to the total number of connections. This is called degree centrality or proximity. It expresses the dis- tance of an individual in the network from other individuals (Freeman, L.C. 1979). Degree centrality (CD) – the activity of an actor in a network is measured by the number of other actors directly connected to it, i.e., the degree of the actor (Bolland, J.M. 1988). Hub (di) – a concept from graph theory, the apex of a network, a point that is connected to other actors in the network (Kőnig, D. 1936). Edge (Li) – a segment connecting the verti- ces forming a graph, each edge running be- tween two vertices (Kőnig, D. 1936). Methodology The methodology developed on the basis of the research questions was tested on an exist- ing data set. The survey was carried out by the Department of Tourism of the Faculty of Sciences, University of Pécs, during the last active tourism peak seasons (in 2018 and 2019 from May to September) before the pandemic, and aimed to map the travel and consump- tion habits of tourists arriving at the South Transdanubian tourist region. The survey was carried out in collaboration with the authors of this study (Gabriella Nod project coordina- tor, and Antal Aubert project manager). The primary data collection was done on paper through a field quantitative survey (in the form of a questionnaire, using assistants to help filling out). The representativeness was based on the spatial distribution, taking into account the data on tourist arrivals previously published by the Hungarian Central Statisti- cal Office (KSH) in 2018. Data processing was performed on a sample of N = 430 items. Sampling was done by a field survey with a Pécs focus. The Baranya County seat accounts for 16 percent of the sample, Szekszárd, Kaposvár and the Harkány– Villány–Siklós triangle have a significant share, similar to the KSH (2018) data. 7.9 percent of respondents were foreigners, 10 percent came to the region from Budapest, and 27.9 percent were travelling within South Transdanubia during the survey. The male/ female ratio among the respondents was 5.5 to 4.5. The top 3 travel motivations were relaxa- tion/regeneration (65.35% of cases), city visits (44.19%) and hiking in nature (25.58%). Data on the accommodation used by respondents, des- tinations visited, travel motivation and general demographic factors influencing the decision were processed. In the sample area, i.e., the South Trans- danubian tourism region, 529,384 guests stayed in commercial accommodations (provid- ing business accommodation and meeting the legal requirements: minimum 5 rooms or 10 beds (Government Decree 239/2009 [X. 20.]) in 2019 (KSH, 2020). Prior to the introduction of the data supply system in 2020, only the guest flows of commercial accommodation were included in the statistics, so we had only estimates of the actual guest night numbers, which is why “invisible or hidden tourism” is an important research topic (Gonda, T. Nod, G. and Aubert, A. Hungarian Geographical Bulletin 71 (2022) (3) 287–299.290 et al. 2018; Michalkó, G. and Ilyés, N. 2020). The big data database to be built from 2020 onwards will now allow tracking guest move- ments and drawing conclusions for the nar- row section of the profession that has access to this data. In the absence of access to data, similar sampling could help track tourists’ spatial movements and improve the available measurement tools and methodology. Mobility map analysis The first part of the research aims to model the tourist movements of tourists within a sample area (region) as a coherent network, and then to determine the position of each destination within the region in terms of the direction and frequency of movements. The network science literature most commonly uses centrality and prestige analysis for position analysis, the for- mer being applied to undirected and the latter to directed networks (Albert, R. and Barabási, A.-L. 2002; Letenyei, L. 2006; Tiszberger, M. 2015; Gao, C. et al. 2022). In the present study, the touristic movements of tourists within the region represent the interconnections of the network, the settlements where tourists used accommodation services became the hubs (di) of the network, the links between them, i.e., the tourist movements, are the edges (Li). The settlement providing accommodation is there- fore the hub and the settlements visited from there form the other elements of the network, i.e., there were tourist movements from hub di to a settlement with k elements. At this stage of the analysis, the network should therefore be treated as a directed network. The cartographic representation of tourist movements was done in QGis 3.10 software and the data analysis (determination of the touristic position of the destination, analysis of the fre- quency of contacts) was done in Microsoft Excel. Focal point testing in the network One possible way to characterize the posi- tions of settlements based on the data is to use degree centrality (CD), where analysis assumes that the degree (i.e., the number of other actors directly connected to it) is a good measure of the activity of an actor, following Letenyei, L. (2006). CD(ni) = d(ni) = ∑ jxij , where d(ni) is the degree of operator i, i.e., the sum of the values in row i of the matrix (Letenyei, L. 2006). The degree number of the settlements d(ni), indicates the number of destinations that send visitors to the municipality. Since the indicator depends on the size of the network, for comparability this number must be di- vided by the maximum value of the network, which is g–1 (if it is connected to all other actors), where g is the number of members in the network, i.e., the number of settlements in the network (in the case of the study: 120). C’D(ni) = d(ni) / (g–1), where d(ni) is the degree of the operator i and g is the number of members in the network (Tiszberger, M. 2015). Let k denote the number of members of the subnetwork formed by di, i.e., all the settle- ments that are connected to di (either as send- ing or receiving parties). Fd(ni) denotes the frequency of connections made. Li denotes the number of connections realized, i.e., all connections made to di, whether outward or inward from di. Ni denotes the number of tourists staying in and visitors to a giv- en settlement within the sample (N = 430). Fd(ni), or frequency of degree, denotes the total number of movements to di from the sending settlements d(ni). Total tourism movements between sending settle- ments and designated settlements as a proportion of potential movements This requires the ratio of total number of tourism movements (Fd(ni)) from the send- ing settlements (d(ni)) to the total number of (1) (2) 291Nod, G. and Aubert, A. Hungarian Geographical Bulletin 71 (2022) (3) 287–299. possible movements to the designated set- tlement (di). In other words, the percentage of visitors staying in the sending settlement who visited the designated settlement: Number of guests staying in the sending settlements / Total number of tourist move- ments to the selected settlement = ∑ (number of guests) d(ni) / Fd(ni). Positioning of settlements according to the pro- pensity of tourists to mobilise The next part of the study focuses on the market positioning of each destination, using data on accommodation and destinations vis- ited from the survey responses. The remain- der of the study also uses the previously de- scribed degree number d(ni), i.e., the number of other settlements that send guests directly to the surveyed settlement. Suppose that by n number of guests stay at settlement di who visit settlement t1, t2 ... tn. We can determine the percentage of the n number of guests staying in settlement di who visit set- tlements t1, t2 ... tn and vice versa. This allows us to measure both an inward and an outward networking. These indicators serve to define the positioning of destinations, which can also optimise the structure of co-operations and destination management. Once the database has been sorted, the study determines the number of guests staying in each settlement (N(gi) number of guests) and the number of settlements visited by guests during their stay in the region (kout), which also indicates the extent of mobility of the settlement. The average number of settle- ments visited by a guest gives the mobility propensity of a settlement (Mw, willingness to be mobile). The “popularity” of a settlement is further measured by the degree number of the settlement d(ni) and the degree frequency Fd(ni). The former represents the number of sending settlements and the latter the num- ber of guests from the sending settlement. Since each settlement has a mobility propen- sity score and a popularity score, a k-means calculation can be used to determine the po- sition of the settlements, i.e., which cluster is closest to the central value of the cluster based on the position of the two scores (Tan, P.-N. et al. 2006): where the coordinates of the mean of each cluster are given as criteria. Analysis of findings The performance of some tourist destina- tions is determined by the size of the send- ing area, so the results are presented first by plotting the movements of visitors to the region between their place of residence and their destination of choice in the region in vector form (Figure 1). (For readability of the map, only movements of domestic tourists are displayed.) In the case of the present study area, the spatially representative survey describes a strong intra-regional movement, with 13.0 percent of the respondents living in Baranya county, 8.7 percent in Somogy county and 6.3 percent in Tolna county (counties of South Transdanubia). A significant proportion of foreign visitors to South Transdanubia come from the capital city Budapest (10.3%) and Pest county (2.6%), Fejér county (8.7%), and the proportion of foreign tourists not shown in the map (8.2%) is also significant. (Given the history of the region and the visiting habits of the expatriate German and Swabian popula- tion still living here, it is not surprising that a significant proportion of foreign visitors are Austrian and German.) Characteristics of the network The analysis of a mobility map helps to de- termine the position of a tourist destination within the region, i.e., whether it is a desti- nation in its own right or whether it offers a complementary service to the region’s tourism (3) ROOT((“Popularity”-INDEX(BLOCK(Cluster mean X,Y))^2 + (Mw-INDEX(BLOCK (Cluster mean X,Y))^2, Nod, G. and Aubert, A. Hungarian Geographical Bulletin 71 (2022) (3) 287–299.292 offer. In total, tourists surveyed used accom- modation services in 60 municipalities and visited 98 settlements with touristic intentions. As the total number of elements of the net- work (accommodation + visited destination) is 120 settlements, corrected by the number of settlements that are both accommodations and destinations, it is not possible to present the full sample, i.e., all the contacts of all the settle- ments, due to space limitations, but the study highlights some of them that demonstrate the applied research methodology. The population size of the municipalities was defined as a cri- terion, with the lower limit being the level of small urban municipality (> 10,000 inhabitants). In Baranya, Tolna and Somogy counties, there are 14 municipalities above this population, 10 of which are statistically assessable on the basis of the survey. This group is completed by Villány and Harkány (Table 1), which do not meet the population criterion but are important in the region because of their small-town status and tourist offer. Although Siófok corresponds to a medi- um-sized city in terms of population and was named by the respondents as a preferred destination in the open-ended questions, as a settlement on the shore of Lake Balaton it was part of the priority tourism development area (Government Decree 429/2016 [XII. 15.]), which means that its development opportu- nities and access to funding sources differ significantly from other typical settlements in South Transdanubia, which is why it was not included in the study. Degree centrality in the network After the data collected during the sampling were digitized and sorted into a database, the data of the 12 settlements to be analysed were selected: the number of respondents using accommodation services in the settlement (the number of guests staying in the settle- ment), and the number of sending settlements (the degree of settlement), i.e., from which settlements the guests staying there came to the settlement under study (Table 2). It is important to note that the data in the table Fig. 1. Tourist arrivals in the South Transdanubian tourist region, 2018. Source: Survey and editing by the authors. 293Nod, G. and Aubert, A. Hungarian Geographical Bulletin 71 (2022) (3) 287–299. only include data from the sample; the sta- tistics shows that these municipalities have proportionally higher guest flows. However, since the sample was spatially representative, the approach is relevant for the methodology. To interpret the data in Table 2, let us take Pécs as an example. In the case of Pécs, 27 other settlements were visited by guests staying here (111 persons) and 35 other ac- commodation establishments (settlements) sent tourists to the county seat. The corrected degree-centrality of the settlement in the to- tal network is 0.29, that is 29 percent of all the settlements in the network are linked to Pécs. Our Pécs-centred network is connected to 45 peaks (each settlement with which Pécs is connected is represented as a peak in the network, whether it is a sending or receiving municipality or both). The adjusted degree- centred value is therefore a good measure of the centrality of a settlement in the net- work, and, where appropriate, its position in the tourism market. The high networking of Pécs as a regional centre is not surpris- ing, while the higher values of Szekszárd, Mohács, Harkány and Villány are associ- ated with a strong local tourism offer. Next, the share of total tourism movements from the sending settlements to Pécs is presented in relation to the potential movements. That is, the percentage of visitors staying in the sending settlement who actually visit Pécs. In the case of Pécs, the rate is 44.1 percent, i.e., almost half of all possible movements to Pécs are made in the Pécs-centred network. A significant value is also seen in the case of Harkány and Villány, where one in four of the guests staying in the sending settlements is sure to visit the settlement that is the centre of the network (Harkány or Villány, as the case may be). The value indicates a kind of likeliness of the proportion of the guest flows in the vicinity of a given settlement within the actual guest flows of that settlement. This value can therefore be used for forecasting and as a tool for strategic planning. The value Li represents the total tourist flows to Pécs, i.e., the number of tourists staying here and the number of visitors coming for one day in the sample. Positioning of settlements according to the mobility propensity of tourists staying in them Of the 430 guests surveyed in the sample, 395 had used accommodation services during their stay in the region. The average number of nights spent by a guest in the region was 4.1 (standard deviation being 3.09). In total, respondents used accommodations in 62 set- tlements and visited 101 settlements for tour- ism purposes. Guests staying in one settle- ment moved to an average of 5.34 additional Table 1. Municipalities in South Transdanubia included in the study Town Local population, persons County Type of settlement Pécs 144,188 Baranya Big city Kaposvár Szekszárd Komló 61,920 32,156 22,832 Somogy Tolna Baranya Medium-sized cities Paks Dombóvár Mohács Bonyhád Tolna Szigetvár Harkány Villány 18,788 17,995 17,143 12,982 10,987 10,545 4,632 2,282 Tolna Tolna Baranya Tolna Tolna Baranya Baranya Baranya Towns Source: Own editing based on 2019 KSH data. Nod, G. and Aubert, A. Hungarian Geographical Bulletin 71 (2022) (3) 287–299.294 settlements during their stay. A settlement was visited by staying guests from another 3.14 settlements, on the average. Based on the number of guests staying in a settlement N(gi) and the number of settle- ments they visited during their stay in the region (kout), we obtain the average number of settlements visited by a guest, which shows the mobility propensity of guests staying in a settlement (Mw,) (Figure 2). Together, the number of tourists staying in a municipal- ity and the number of guests visiting it in the course of the query give the “popularity” indicator of the settlement, which is catego- rised by colours in Figure 3. The mobility propensity score ranges from 0.3 to 8.0 for the present study, where three broad groups are typically distinguished. The value below 1.0 is typical of those settlements where one or two guests’ opinions were col- lected and the number of other settlements visited by them is low; the group’s use of ac- commodation is characterised by more cost- effective solutions (33.3% stayed with a rela- tive/friend, 23.8% in a boarding house, 14.3% in a holiday home). Bikal is an exception in the group, with higher accommodation ex- penditure (87.5% of Bikal guests stayed in a hotel), but Bikal and its medieval-style experi- ence facility offer a complete stay of several days, which explains the low mobility. The next groups are typically medium- sized towns and large cities, with high num- bers of guests and lower mobility willing- ness, explained by the complex and multi- functional tourism offer of the destinations. Smaller municipalities also scored high in terms of mobility, where, as in the first cat- egory, the number of guests staying and the expenditure on accommodation are typically low (46.7% staying with friends/relatives, 26.7% in campsites and 20–20% in rural ac- commodation or boarding houses) but where mobility is high. In the light of the data, this group is characterised by a longer length of stay (6.6 nights on average). The mobility propensity values, comple- mented by the number of guests N(gi), the number of degrees d(ni) and the frequency of Ta bl e 2 . M et ric s o f t he n et w or ks in th e s ur ve ye d m un ic ip al iti es To w n N um be r o f gu es ts s ta yi ng in th e se ttl em en t N um be r o f de st in at io n vi si te d, p cs D eg re e ce nt ra lit y, d (n i) D eg re e of th e se ttl em en t, C ’D (n i) N um be r o f n et w or k ite m s, p cs Fr eq ue nc y of de gr ee , F d( n i) Fd (n i) co m pa re d to a ll po ss ib le m ov em en ts , % Pa ss in g ed ge s Li , p cs Pé cs K ap os vá r Sz ek sz ár d K om ló Pa ks M oh ác s D om bó vá r Bo ny há d To ln a Sz ig et vá r H ar ká ny V ill án y 11 1 11 35 10 5 11 20 6 2 3 36 16 27 10 20 13 5 9 11 4 3 5 18 8 35 7 23 3 5 18 7 8 1 11 13 19 0. 29 0. 06 0. 19 0. 03 0. 04 0. 15 0. 06 0. 07 0. 01 0. 09 0. 11 0. 16 45 13 35 13 7 22 14 10 4 15 25 22 11 6 18 33 8 9 33 10 12 2 38 56 68 41 .1 9. 7 16 .7 4. 8 16 .7 12 .2 6. 3 7. 0 5. 7 14 .4 30 .4 24 .6 22 7 29 68 18 14 44 30 18 4 41 92 84 So ur ce : O w n su rv ey . 295Nod, G. and Aubert, A. Hungarian Geographical Bulletin 71 (2022) (3) 287–299. Fig. 2. The surveyed settlements of the South Transdanubian tourism region according to the mobility propensity of the visitors staying there. Source: Survey and editing by the authors. Fig. 3. The popularity of the surveyed settlements in the South Transdanubian tourism region in terms of the number of guests and visitors. Source: Survey and editing by the authors. Nod, G. and Aubert, A. Hungarian Geographical Bulletin 71 (2022) (3) 287–299.296 degrees Fd(ni), which help to measure the “popularity” of the settlements, divide settle- ments into six categories: less preferred mu- nicipality with high propensity to move out; less preferred municipality with low propen- sity to move out; medium preferred munici- pality with higher propensity to move out; medium preferred municipality with lower propensity to move out; preferred municipal- ity with high propensity to move out; and preferred municipality with low propensity to move out. The k-means calculation used in the cluster analysis, based on the position of each municipality according to two values (popularity value and mobility propensity), placed the municipalities in the group closest to the central value of each cluster (Table 3). The municipalities receiving each cluster value and their spatial distribution within the sample area are illustrated in Figure 4. Figure 4 distinguishes six clusters. As in Asero, V. et al. (2016), the categories can be generalised and applied to the development of regional or (depending on the available data) national strategies. The number of visi- tors staying in and visiting a municipality is an indicator of its popularity, while the mobil- ity propensity measures the tourist retention capacity of the municipality, the lower the mo- bility, the stronger the tourist retention capac- ity of the municipality. At the same time, the multiplier effect of tourism may be stronger in the vicinity of settlements with high mobility, i.e., settlements in the vicinity are more likely to experience an increase in visitor numbers. Based on the mobility propensity scores of vis- itors to a settlement, another grouping can also be applied, similar to the categories known from the literature on settlement geography (Pirisi, G. and Trócsányi, A. 2019): – Independent settlement from which the resi- dent guest does not move or moves only slightly. These are low mobility settlements with a group-specific or broad touristic offer. – A “sleeping” settlement, with a favourable accommodation offer for the visitor, but lit- tle or no other tourism offer; characterised by a high mobility propensity. – A cooperative settlement with its own touris- tic offer, but its strength lies at the regional level, where it creates, together with other municipalities in the region, an attractive touristic offer. Conclusions The study applies two approaches based on the methodological tools of network theory. To support the hypothesis formulated in the introduction (the role of individual destinations in the supply market of the region can be measured by the tourist flows across destinations), it primarily models the tourist flows within the region as a coherent network. This is demonstrated in chapter 3.2 by presenting the characteristics of the network (degree number and centrality), i.e., by answering question Q1 (Can tourism movements in the region be interpreted as a coherent network?) and Q2 (Can the charac- teristics of the network (centrality, degree) be used to depict the position of each destina- tion within the region?). The degree number and the centrality of the settlements provide a Table 3. Clusters defined by k-means Cluster number Popularity, persons Mobility, Mw settlement/person Cluster characteristics 1st 2nd 3rd 4th 5th 6th 1.00 1.45 5.22 5.91 28.50 109.80 7.33 2.24 3.05 1.46 1.77 1.82 Less preferred with high propensity to move out Less preferred with low propensity to move out Medium preferred with higher propensity to move out Medium preferred with lower propensity to move out Preferred with high propensity to move out Preferred with low propensity to move out Source: Own survey. 297Nod, G. and Aubert, A. Hungarian Geographical Bulletin 71 (2022) (3) 287–299. good measure of their position in the region’s tourism offer, and their links, which can be statistically demonstrated by analysing the data, can also serve as a guide for cooperation. Finally, research question Q3 would char- acterise a given settlement on the basis of the mobility propensity of the guests stay- ing in a destination. To demonstrate this, the study calculates a mobility propensity score for each settlement where respondents used accommodation services at the time of the survey. It then calculates the popularity of the settlement based on the number of guests staying there, the degree number of the settle- ment and the frequency of the degree num- ber. Finally, clusters are formed based on the indicators of popularity of the settlement and the mobility propensity scores. The resulting clusters are geographically independent and can be applied in practice anywhere. They are of strategic importance as they can be used as logical categories in development plans. The weakness of the study is that it does not use a complete regional database, only a small sample. This flaw can be overcome by accessing the above-mentioned national statistical database under construction. This type of data collection is also an internation- ally known practice, and so it is not confined within the country’s borders. It is innovative in the sense that few studies have so far dealt with network analysis of tourism movements at international level. It identified network characteristics based on the tourism move- ments studied and presented a possible clustering of tourist destinations based on the popularity of the locality and the mobil- ity propensity of the visitors staying there. Fig. 4. Clusters based on the mobility propensity and popularity of the studied settlements in South Transdanubia. Source: Own survey, own editing. Nod, G. and Aubert, A. Hungarian Geographical Bulletin 71 (2022) (3) 287–299.298 Acknowledgements: The PhD student, as project coor- dinator is grateful to the Márton Áron Special College of the Eötvös Loránd University. The research was supported by the ÚNKP-21-3 New National Excellence program of the Ministry for Innovation and Technology from the source of the National Reseach, Development and Innovation Fund. REFERENCES Ahas, R. and Mark, U. 2005. Location based services – New challenges for planning and public admin- istration? Futures 37. (6): 547–561. Albert, R. and Barabási, A.-L. 2002. Statistical me- chanics of complex networks. Reviews of Modern Physics 74. 47–97. Aubert, A., Berghauer, S., Beták, N., Mókusné Pálfi, A. and Nod, G. 2021. Turisztikai szervezetek online marketing tevékenységének vizsgálata: a szlovák-magyar-ukrán (kárpátaljai) példa (Survey of online marketing activities of tourism organisations: a Slovakian-Hungarian-Ukrainian [Transcarpathian] example). In Interregionális kutatási eredmények a turizmusban: örökségvédelem, intézményrendszer, helyi termékek (Hun-Au-Slo- Ukr). Eds.: Mókusné Pálfi, A. and Nod, G., Pécs, University of Pécs, Department of Tourism, 70–90. Asero, V., Gozzo, S. and Tomaselli, V. 2016. Building tourism networks through tourist mobility. Journal of Travel Research 55. (6): 751–763. Available at https://doi.org/10.1177/0047287515569777 Barabási, A.-L. 2016. Network Science. Cambridge, UK, Cambridge University Press. Bernecker, P. 1962. Grundlagenlehre des Fremdenverkehrs. Wien, Österreich, Österreichchischer Gewerbeverlag. Bolland, J.M. 1988. Sorting out centrality: An analysis of the performance of four centrality models in real and simulated networks. Social Networks 10. (3): 233–253. Butler, R.W. 1980. The concept of the tourist area life-cycle of evolution: implications for manage- ment of resources. Canadian Geographer 24. (1): 5–12. Casanueva, C., Gallego, Á. and García-Sánchez, M-R. 2014. Social network analysis in tourism. Current Issues in Tourism 19. (2): 1190–1209. Available at https://doi.org/10.1080/13683500.2014.990422 Csapó, J. and M. Császár, Zs. 2021. A Magyarországon tanuló külföldi egyetemi hallgatók turisztikai fogyasztási szokásainak elsődleges elemzése (A primary analysis of tourism consumption habits of international university students studying in Hungary). Turizmus Bulletin 21. (2): 15–23. Available at https://doi.org/10.14267/TURBULL.2021v21n2.2 D’Agata, R., Gozzo, S. and Tomaselli, V. 2013. Network analysis approach to map tourism mobility. Quality and Quantity 47. (6): 3167–3184. Available at https://doi.org/10.1007/s11135-012-9710-7 Díez-Díaz, F., González-Rodríguez, M. and Vidau, A. 2007. An accessible and collaborative tourist guide based on augmented reality and mobile devices. In Universal Access in Human-Computer Interaction. Ambient Interaction. UAHCI 2007. Lecture Notes in Computer Science, Vol. 4555. Ed.: Stephanidis, C., Berlin–Heidelberg, Springer. Available at https:// doi.org/10.1007/978-3-540-73281-5_38 Erdős, P. and Rényi, A. 1960. On the evolution of random graphs. Magyar Tudományos Akadémia Matematikai Kutató Intézet Közleményei 5. 17–61. Euler, L. 1741. Solutio problematis ad geometriam situs pertinentis. Euler Archive – All Works, 53. Available at https://scholarlycommons.pacific.edu/ euler-works/53 Freeman, L.C. 1979. Centrality in social networks con- ceptual clarification. Social Networks 1. (3): 215–239. Gao, C., Zhong, Y., Mensah, I.A., Tao, S. and He, Y. 2022. Spatio-temporal impact of global migration on carbon transfers based on complex network and stepwise regression analysis. Sustainability 14. (844): 1–19. Available at https://doi.org/10.3390/ su14020844 Gonda, T., Raffay, Z. and Csóka, L. 2018. Orfű rejtett konvencionális turizmusa (The hidden conventional tourism of Orfű). Turisztikai és Vidékfejlesztési Tanulmányok 3. (1): 4–19. Available at https://journals.lib.pte.hu/index.php/tvt/article/ download/4368/4100/8264 Government Decree 239/2009 (X. 20.) on the “detailed conditions for pursuing accommodation service activities and the procedure for issuing accommo- dation licenses”. Government Decree 429/2016. (XII. 15.) on the „pri- oritized tourist development areas”. Government Decree 429/2020 (IX. 14.) on the “defini- tion of tourist regions”. Hungarian Tourism Agency 2021. Tourist Regions. Available at https://mtu.gov.hu/cikkek/11-turisz- tikai-terseg-terkep-es-telepuleslista-1918 (accessed on 21/11/2021) (in Hungarian) Kotler. P. 1967. Marketing Management, Analysis, Planning and Control. Upper Saddle River, NJ, USA, Prentice-Hall. Kőnig, D. 1936. Theorie der endlichen und unend- lichen graphen. Kombinatorische Topologie der S t r e ckenkomplexe . Le ipz ig , Akademische Verlagsgesellschaft M. B. H. KSH 2018, 2020, 2021. Hungarian Central Statistical Office. Regional breakdown of tourism data of com- mercial accommodation establishments. Available at https://www.ksh.hu/stadat_eves_4_5 (accessed on 10/12/21). Letenyei, L. 2006. Településkutatás (Settlement re- search). Tér és Társadalom 20. (2): 143–156. Available at https://doi.org/10.17649/TET.20.2.1911 299Nod, G. and Aubert, A. Hungarian Geographical Bulletin 71 (2022) (3) 287–299. Madarász, E. and Papp, Zs. 2013. Delimiting the “Balaton Riviera”. Tourist destination by using network analysis. Hungarian Geographical Bulletin 62. (3): 289–312. Available at https://ojs.mtak.hu/ index.php/hungeobull/article/view/2963 Mas-Colell, A., Whinston, M.D. and Green, J.R. 1995. Microeconomic Theory. Oxford, UK, Oxford University Press. Michalkó, G. and Ilyés, N. 2020. A nemkonvencioná- lis turisztikai mobilitás magyarországi sajátosságai, különös tekintettel a turizmusból származó bev- ételek növelésének lehetőségére (The specificities of unconventional tourism mobility in Hungary, with a special focus on the potential for increasing tourism revenues). Turizmus Bulletin 20. Special number. 6–18. Doi: 10.14267/turbull. 2020v20n4.1 National Tourism Development Strategy 2030. Budapest, Magyar Turisztikai Ügynökség. Available at https://mtu.gov.hu/documents/prod/NTS2030_ Turizmus2.0-Strategia.pdf (in Hungarian) Nod, G., Pálfi, A. and Aubert, A. 2019. Az egyéni utazásszervezés térnyerése a digitális társadalom- ban (Raising of the individual organizations of passenger travels in the digital society). In Turizmus, fogyasztás, generációk. II. Nemzetközi Turizmus Konferencia: Tanulmánykötet. Eds.; Csapó, J. et al., Pécs, University of Pécs, 161–171. Pécsek, B. 2015. The role of slow design elements in managing tourist flow on the example of Bruges, Belgium. Hungarian Geographical Bulletin 64. (2): 143–154. Available at https://doi.org/10.15201/ hungeobull.64.2.5 Pirisi, G. and Trócsányi, A. (eds.) 2019. Fejezetek a társadalomföldrajz világából (Chapters from the world of social geography). PTE TTK Földrajzi és Földtudományi Intézet. Pécs, Publikon Kiadó. Piskóti, I. 2007. Desztináció-menedzsment lépésről-lépésre (Destination management step-by-step). Eger- Miskolc, NORDA. Scott, N., Baggio, R. and Cooper, C. 2008a. Network Analysis and Tourism. From Theory to Practice. Aspect of Tourism series. Bristol, UK, Channel View Publications. Scott, N., Cooper, C. and Baggio, R. 2008b. Destination networks: Four Australian cases. Annals of Tourism Research 35. (1): 169–188. Scott, N., Cooper, C. and Baggio, R. 2009. Use of Network Analysis in Tourism Research. First Edition. Aspect of Tourism series. Bristol, UK, Channel View Publications. Spinney, J.E. 2003. Mobile positioning and LBS applications. Geography 88. (4): 256–265. Sziva, I. 2014. The role of cooperative strategies in destinations’ competitiveness in a turbulent environment. Journal of Tourism and Hospitality Management 2. (3): 97–105. Tan, P-N., Steinbach, M. and Kumar, V. 2006. Introduction to Data Mining. 1st Edition. New Jersey, Pearson Education, Inc. Available at https://gyires. inf.unideb.hu/KMITT/a04/ch08s02.html Telbisz, T., Šulc, I., Mari, L. and Radeljak Kaufmann, P. 2022. Attitudes and preferences of visitors of Krka National Park, Croatia. Hungarian Geographical Bulletin 71. (2): 117–132. Available at https://doi. org/10.15201/hungeobull.71.2.2 Tiszberger, M. 2015. A hálózatkutatás módszer- tani vizsgálati lehetőségei – szakirodalmi össze- foglalás (Methodological options for network research – a literature review). Pécs, University of Pécs. Available at https://ktk.pte.hu/sites/ ktk.pte.hu/files/images/szervezet/intezetek/ mti/tiszberger_a_halozatkutatas_modszertani_ vizsgalati_lehetosegei_2015.pdf Tóth, G. and Dávid, L. 2009. Az elérhetőség és az ide- genforgalom kapcsolata (The connection between accessibility and tourism). Tér és Társadalom 23. (3): 45–62. Doi: 10.17649/TET.23.3.1256. Watts, D.J. and Strogatz, S.H. 1998. Collective dynamics of ‘small-world’ networks. Nature 393. (6684): 440–442. Doi:10.1038/30918. PMID 9623998. Wind, S. 2015. Simulating and modelling urban flows: Experimenting with tracking technologies in urban design. Arkitektur & Design. A&D Skriftserie Vol. 91. Available at https://vbn.aau.dk/ws/files/214876372/ SUF_booklet_SMALL.pdf Woodside, A. and Lysonski, S. 1989. A general model of traveller destination choice. Journal of Travel Research 17. (4): 8–14. Zheng, W., Huang, X. and Li, Y. 2017. Understanding the tourist mobility using GPS: Where is the next place? Tourism Management 59. (4): 267–280. Available at https://doi.org/10.1016/j.tourman.2016.08.009 Nod, G. and Aubert, A. Hungarian Geographical Bulletin 71 (2022) (3) 287–299.300