Tourism Management 52 (2016) 30e43

Contents lists available at ScienceDirect

Tourism Management

journal homepage: www.elsevier.com/locate/tourman

Case study Understanding tourist space at a historic site through space syntax analysis: The case of Gulangyu,

* Yuan Li a, b, , Longzhu Xiao a,YuYec, Wangtu Xu a, Andrew Law d a School of Architecture and Civil Engineering, University, 422 South Siming Road, 361005 Xiamen, China b College of Architecture and Urban Planning, Tongji University, 1239 Siping Road, 200092 Shanghai, China c Faculty of Architecture, The University of Hong Kong, Knowles Building, Pokfulam Road, Hong Kong d School of Architecture, Planning and Landscape, Newcastle University, Claremont Tower, Claremont Road, Newcastle upon Tyne NE1 7RU, United Kingdom highlights

We examine space syntax analysis as a reflection of general morphological structure. We study the effect of street network integration on tourist preferences. We evaluate the planned tourist sites based on space syntax analysis. We provides a visualized and quantitative approach to the field of tourist management via the application of GIS. article info abstract

Article history: Although tourist development appears to be an effective approach for the further development of his- Received 5 November 2014 toric towns, researchers have often contended that a balance is difficult to maintain between protection Received in revised form and development. This article applies space syntax analysis to provide an understanding of tourist space 29 May 2015 at historic sites on Gulangyu Island. The relationship between street network integration and the urban Accepted 13 June 2015 fabric as well as tourist preferences collated from data mining are explored and verified. Then, results of Available online xxx the space syntax analysis along with the proposed historic site plan are compared to provide suggestions in plan evaluation. This research provides a visualized and quantitative approach to the field of tourist Keywords: Tourism management management via the application of Geographic Information System (GIS) to urban morphological fea- Space syntax tures, tourist preferences and results gathered by way of planning practice. The effectiveness and limi- Urban morphology tations of this proposed new approach are also debated and substantiated. Tourist preferences © 2015 Elsevier Ltd. All rights reserved. Gulangyu

1. Introduction historic sites is becoming more and more pressing. Nevertheless, recent advances in spatial analytical tools bring The protection of historic heritage and the sustainable reorga- new opportunities for planning tourist spaces more appropriately nization of tourist space has become a common agenda for ur- than before. According to the socio-spatial dialect, spatial ar- banists (Hayllar, Griffin, & Edwards, 2008; McKercher, Ho, & du rangements can reflect certain social structures and cultural con- Cros, 2005; Wang & Bramwell, 2012). However, the conflict be- notations, and vice versa (Soja, 1980). In other words, by tween heritage protection and tourism development is also glar- understanding the inherent morphological structure and behav- ingly evident, especially for developing countries (Li, Wu, & Cai, ioral characteristics of a space, more sustainable and well organized 2008). Due to the pressures of a rising population and the pursuit tourist development approaches be easily achieved. of local economic goals, the need to better organize tourist space at The main method utilized in urban morphological and behavior research is space syntax. Over the years, space syntax analysis has contributed to a greater understanding of the spatial configuration * Corresponding author. School of Architecture and Civil Engineering, Xiamen of street and road networks, and how these kinds of configuration University, 422 South Siming Road, 361005 Xiamen, China. affect movement flows, the location of economic activities, and the E-mail addresses: [email protected] (Y. Li), [email protected] (L. Xiao), numerical levels of street life (Hillier, Penn, Hanson, Grajewski, & [email protected] (Y. Ye), [email protected] (W. Xu), andrew.law@newcastle. ac.uk (A. Law). Xu, 1993; Hillier, 1996b; Karimi, 2000, 2012). With the help of http://dx.doi.org/10.1016/j.tourman.2015.06.008 0261-5177/© 2015 Elsevier Ltd. All rights reserved. Y. Li et al. / Tourism Management 52 (2016) 30e43 31 computer software, like Depthmap or sDNA, it is possible to visu- transformations. During this period then, Local authorities carried alize the spatial hierarchy of street networks so that we can develop out tourist oriented development policies. The peripheral areas a picture of a particular urban morphological structure, in order to within the Longtou area (especially near the docks) were developed predict pedestrian flows (Griffiths, Jones, Vaughan, & Haklay, 2010; into tourist service centers. During this process factory districts Ye & van Nes, 2013, 2014a, 2014b). In this paper, traditional were moved out to the mainland. morphological analysis focusing on urban fabrics will be firstly As a well renowned tourist destination located near China's applied to test the accuracy of the space syntax approach. Real time southeast coast, Gulangyu is currently facing a dramatic conflict tourist movement behavior recorded through large data mining of between booming tourist development and a crucial need to pro- cell phone locations is utilized to provide another perspective in tect the historic heritage of the island. Since the 1990s, local au- verifying the understanding of morphological features and tourist thorities have pursued pro-tourist development policies and as a preferences quantitatively. Tracing tourist preferences through cell result commentators have contended that both the existing com- phone or GPS technologies is a new technique which has arisen in munity culture and the attractiveness of the island have increas- recent years (Ratti, Williams, Frenchman, & Pulselli, 2006; Shoval, ingly been under threat (Lin, 2010; Yu & Liu, 2011). Statistical data 2008). As a technique able to record the real time movement of shows that Gulangyu received more than 11,360,000 tourists in people, it brings much new potential for tourist management. 2013, and the number is expected to hit 13 million by the end of The following paper is composed of five distinct parts. Following 2014 (Administrative Committee of Gulangyu, 2014). During the the introduction, Section 2 illustrates the research methodology. In May holidays of 2014, the number of tourists visiting the Island per this section we examine the case study and the analysis methods. day exceeded 70,000 (Fig. 2). We also discuss our hypothesis and broader issues concerning our In this context, Gulangyu has been included in the National research design. Section 3 examines and explores our empirical Preparatory Directory for World Cultural Heritage, which proposes analysis and discussions. In this section, specifically we unpack our that tourist development must give priority to historic heritage time series analysis of the morphological structure of the research protection. Apparently, as Yu and Liu (2011) has claimed, for case studies, using the space syntax method; after this process we Gulangyu to safe guard the community culture and heritage of the then overlay the results of the space syntax analysis on a traditional Island, planners and designers need to develop a deep and rich morphological analysis. Then, we investigate the overlaying anal- understanding of Gulangyu's inherent morphological structure and ysis between street network configurations resulting from space how this structure affect tourist behavior. syntax and tourist preferences. Finally, suggestions for optimizing tourist space organization and approaches for sustainable historic 2.2. Analysis methods conservation will be discussed. In Sections 4 and 5, the strengths and limitations of the research and the conclusions will be Space syntax is both a theory of urban planning and design and a reviewed respectively. software-based technology. As a theory, space syntax differs from classical urban morphology because it focuses on open space sys- 2. Methodology tems to pursue a form of spatial representation (Hillier & Hanson, 1984). Space syntax theorists call sets of simulated relations be- 2.1. Research case tween spaces as “configuration” or “accessibility” and these terms are used as ways of describing what they understand to be the As an island located to the east of the coast of Xiamen (in intrinsic properties of space (Hillier et al., 1993; Hillier, 1996a, province in southern China), Gulangyu is around 2 km2 (0.77 sq. mi) 1999a). Hillier, therefore argues that the form of space, its configu- in area. It is home to about 20,000 people and is an increasingly ration, can be regarded as the main determinant of urban movement popular domestic tourist destination. The only vehicles permitted flows. One can easily analyze numerical graphs to investigate the on the Island are small electric buggies and electric government relative importance of each space in the whole spatial system. This service vehicles (Gulangyu Island, n.d.). argumentation has been partly verified by indicators of movement As shown in Fig. 1, the evolution of the urban fabric and func- flows (e.g. human behavior records, micro-scale business distribu- tions of Gulangyu has left a historic urban footprint at the site. The tions) in numerous cases during the past ten years. It is possible to primary streets of Gulangyu were systematically formed at the end argue then, that space syntax theory sheds some light on the of the 19th century when many consular officials, priests and development of urban morphological studies from their empirical businessman reached the Island (Hong, 1999; He, 2007). The Nei- and qualitative roots towards a more quantitative based approach. cuo'ao area, where the Municipal Council resided, was the site of Space syntax analytics use three types of distance metrics the earliest aboriginal settlements which were characterized by (Turner, Penn, & Hillier, 2005; Turner, 2007)(Fig. 3): 1) topological mixed functions and scattered layouts. The Tianwei area was (fewest turns paths), 2) geometrical (least angle change paths) and composed of villas and public facilities and is characterized by its 3) metric (shortest physical paths). The instrument then calculates free layout and urban fabric which mirrors the appearance of the the street network configuration at different scales (radii) accord- colonial buildings which were built years later at the end of the ing to the size of the research area. The global scale analysis mea- 19th century (Fig. 1a). In Fig. 1b, the Luqiao-fujian area is shown. sures the average depth of an axial line or segment to all other axial The area is a reflection of construction features lines or segments in whole street networks, while the local scale found during the ‘formation and prosperity’ period (1890se1960s). analysis makes the calculation in a limited radius. In this research, The Luqiao-fujian area is characterized by group layouts, courtyard- local scale analysis is defined as 300 m radius and global scale style buildings, and long quiet streets. The Longtou area which is analysis is defined as 3000 m radius. The DepthmapX developed by marked in dark gray in Fig. 1c is the result of the accumulation of a UCL (University College London) and the measurement of angular fairly complex development process on the Island with its compact segment analysis by metric distance (ASAMeD) inside the software buildings and packed streets. Its naturally generated spatial struc- are applied to our case, so as to measure the spatial configuration of ture provides us with clues so that we can begin to grasp the his- street networks in Gulangyu's morphological structure. The toric development process, cultural diversity and social structure of ASAMeD contains two measures e “closeness” and “betweenness”. ‘ ’ e the Island. The pro-tourist development period (1970s 2010s) AccordingP to Hillier and Iida (2005), closeness (integration) is 1 displayed in Fig. 1c and d illustrates followed historic Cc(Pi) ¼ ( k dik) , where dik refers to the length of a geodesic 32 Y. Li et al. / Tourism Management 52 (2016) 30e43

Fig. 1. Morphological analysis based on urban fabric and functional structure from the end of the 19th century to 2013.

(shortest path) between node Pi and Pk. This equation reflects how research on visitors' behavior and street networks (Griffiths, & close each segment is to all others underP differentP types of dis- Vaughan, Haklay, Emma Jones, 2008; Hillier, 1999b, 2002; tances. Betweenness (choice) is CB(Pi) ¼ j k gjk(pi) ¼ gjk(j < k), Ismail, Chau, & Rahman, 2010). Often this work demonstrates where gjk(pi) is the number of geodesics between node pj and pk that outside visitors prefer to navigate the streets showing high which contain node pi and gjk the number of all geodesics between global integration; while local residents prefer to choose streets pj and pk. This equation reflects how much movement is likely to demonstrating high local integration. Spatial integration effects pass through each segment on trips between all other segments. outside visitors' choices from a city-wide perspective, while the scope of local residents' daily activities are relatively limited. Ac- 2.3. Research design cording to the space syntax glossary (Klarqvist, 1993), 10% of the most integrated spaces are those normally referred as the inte- 2.3.1. Hypothesis: the analysis of the spatial integration of street gration core. In other words, high global integration, (especially the networks will provide a better understanding of tourist space highest 10%) as a morphological feature, decides whether a place is Numerous studies have applied space syntax analysis to easily accessed; while local integrations (especially the highest Y. Li et al. / Tourism Management 52 (2016) 30e43 33

Fig. 2. Tourist sites in Gulangyu during the May holidays of 2014.

10%) determine whether a place can become vital and attractive. 2.3.2. Research questions and analytical framework Turning to our research, places with high global integration might Confronting the direct pressure of increasing tourist numbers be important routes for tourists who are passing by, while low and declining historic culture, the research objective of this paper is global integration would be an obstacle for tourists' arriving at to construct an analytical approach which is able to optimize tourist these sites. Places that manage to balance global and local in- development and historic heritage protection based on the un- tegrations can lead to an overlap of tourists and local inhabitants' derstanding of morphological structures of the site. Three research behavior; these places can easily become tourist sites. If this hy- questions are proposed to fulfill our objectives: pothesis is valid, quantified street-network integrations can reflect tourist preferences from an urban morphological perspective. Our 1) Firstly, how do we understand the general morphological analysis can then be used to evaluate and optimize the organization structure of Gulangyu and can space syntax analysis provide us of tourism. Nevertheless, the effectiveness of our hypothesis still with an accurate illustration of the morphological structure of needs further verification via real empirical case studies which the Island? shall be explored. 2) Secondly, how do we understand the effect of street network integration and how might this network integration affect tourist preferences and/or behavior characteristics? 3) Thirdly, based on the questions above, how might we begin to evaluate and optimize planned historic sites?

To respond to these three research questions we now propose the following analytical framework (Fig. 4). Firstly, urban fabric and functions recorded in the historic maps will be analyzed through both space syntax and traditional morphological analysis. The ex- planations from the space syntax analysis will be compared with traditional morphological analysis and a general understanding of the morphological structure will be generated (as research objec- tive 1). The spatial structure reflected from our space syntax anal- ysis will then be compared with observed tourist flows via a big data approach to understand the effect of street network integra- tion and the influence that this network integration has on tourist preferences (this is research objective 2). Furthermore, evaluations and suggestions for planned historic sites will be constructed in this analysis (this is research objective 3).

3. Analysis and discussions

3.1. Understanding the general spatial structure via space syntax

3.1.1. Space syntax analysis following time series As we have mentioned above, the space syntax analysis in this research will be conducted through an angular segment analysis by metric distance (ASAMeD) in Depthmap software. Moreover, as we have already mentioned, this mode of analysis enables us to add topological and geometrical complexity into purely metric con- siderations so as to achieve a better reflection of the street network configuration (Hillier & Iida, 2005). In particular, the street network integration of Gulangyu is investigated at two scales. The global and local radii are defined as 300 m and 3000 m according to the size of Fig. 3. The topological analysis of space syntax based on axial lines (aþb) with an Gulangyu. We use the ASAMeD method to calculate the street additional consideration of angular weighting (c) (Turner et al., 2005; Turner, 2007). network integration by setting the radii at 300 m and 3000 m 34 Y. Li et al. / Tourism Management 52 (2016) 30e43

Fig. 4. Illustrated cognitive map of our analytical framework in research. respectively. The global and local integration results are illustrated the global level is visible as Fig. 5d shows. This centrality process in Figs. 5 and 6 following a time series. In the figures, the colors are moved toward the direction of the Sanqiutian Jetty and the Longtou used refer to high and low values of integration. Thus red (in web Wharf where tourist development projects were conducted. version) refers to higher values of integration, while blue (in web Concomitantly, when we speak of the local level, as Fig. 6d shows, version) refers to low values of integration. High or low values are the distribution of the integration cores is broadly in line with the then measured through the natural break method widely used in measures that took place in 1984. The subtle difference, however, is GIS, which minimizes the average deviation of each class from its that the local integration cores that formed in the Longtou area and own mean values while maximizing the deviation of each class the Neicuo'ao area were strengthened while the others were from the mean values of other groups. weakened. That is, a global integration core and several local As Fig. 5a shows, a global integration core arose at the end of the integration cores were staggered spatially which has, in turn, 19th century. Spatial centrality is located at the space where resulted in a multicenter pattern in the present. Huangyan Rd and Longtou Rd cross. When referring to the local level, as Fig. 6a shows, two integration cores are formed in the 3.1.2. Discussion I: space syntax analysis results could reflect the Neicuo'ao area where the Municipal Council and Luqiao-fujian area general spatial structure of Gulangyu well lie (specifically where the churches lie). It follows that the global Whether space syntax, as a method focusing on one of the integration core and the local integration core are spatially stag- essential elements of urban morphology e street network e can gered. Thus, as a result of this staggering there is no absolute center provide a general description of whole urban morphological fea- on the Island because of the separation of the global integration tures has been a subject of discussion for a long time. Many core and the local integration core. scholars believe that since street network configurations are the As demonstrated in Fig. 5b, the global integration core migrates result of many morphological elements e like land use, density, etc slightly to the Longtou area which was the geographical center of e space syntax analysis could, to a certain extent, represent whole the island during the ‘formation and prosperity’ period urban morphological features (Peponis, Ross, & Rashid, 1997; (1890se1960s). Notably, the local integration core formed in the Turner, 2003). On the other hand, other commentators have sug- Luqiao-fujian area at the end of the 19th century almost dissolves, gested that it would be better if space syntax analyses are utilized while during this period there developed a new local integration and combined with other morphological investigative features core in the Longtou area where the global integration core was (Ratti, 2004; Ståhle, Marcus, & Karlstrom,€ 2005; Ye & van Nes, located (as Fig. 6b shows). Therefore the Longtou area is the exact 2012). To date, after many debates in the fifth and sixth space center of Gulangyu during this period due to the overlap of the syntax symposiums, certain academics have suggested that space global and local integration cores. In addition, the local integration syntax analysis can explain whole urban morphological features, core which formed in the Neicuo'ao area at the end of 19th century especially in small historic towns. For other cases, like modern shifts marginally south. mega cities, a combination of street networks and other Interestingly though, the global integration core moves back to morphological features would be more appropriate. Therefore, it is its original point as the 1984 integration map in Fig. 5c demon- reasonable to assume that in the research case discussed in this strates. Referring to the local level, Fig. 6c shows several new paper, (a historic town), space syntax itself might be able to work integration cores that have emerged and have dispersed across the as a quantitative description of urban morphological features, island (in addition to the local integration core that existed prior to which can be then applied to research on the organization of this growth). These local integration cores are separated by the tourism. mountains. The absolute center formed during the ‘formation and In order to respond to this assumption and achieve insightful prosperity’ period between 1890s and the1960s dissolved in 1984. understandings of the morphological features of our research case, During the ‘pro-tourist development’ period of the 1970s to the we have combined traditional morphological analysis and space 2010s, a significant outward tendency of the centrality process at syntax analysis (Fig. 7). The historic transformation of urban fabrics, Y. Li et al. / Tourism Management 52 (2016) 30e43 35

Fig. 5. A figure demonstrating an angular segment analysis with metric radius (Rad-3000 m) analysis of Gulangyu from the end of the 19th century to 2013. urban functions and the spatial integration of street networks are Through the analysis above, it is reasonable to say that generally shown together. What has happened in the urban fabrics and urban speaking, contemporary Gulangyu has a decent spatial structure. functions is reflected in the street network integration. Specifically, The newly developed tourist center has high global integration that then, the generation of historic centers like Neicuoao and Longtou, is convenient and beneficial for tourists which, in turn has a rela- and their changing urban functions can be found synchronously in tively higher mean value (761), compared with the global mean both traditional and space syntax analysis. Therefore, it is reason- value in Gulangyu (Fig. 8). Meanwhile, other historic centers, such able to conclude that in the historic towns of Gulangyu, space as the Longtou and Neicuoao areas, obtain quite high local in- syntax analysis can illustrate general morphological features. Based tegrations (5225 in Neicuoao and 4295 in Neicuoao) and relatively on this kind of quantitative description of street network in- high global values. The co-presence of excellent global and local tegrations, much analysis can be made which can then support integration means that tourists can gain easy passage to the site and research into tourist space management and historic building also experience its townscape and history without the uncom- protection; in sum, space syntax analysis, such as the analysis fortableness of overcrowding. developed in this paper, could be of significant use to urban plan- However, despite the good levels of integration, the problem for ners and conservationists. tourist management is that the existing high pressure (mainly from 36 Y. Li et al. / Tourism Management 52 (2016) 30e43

Fig. 6. A figure demonstrating angular segment analysis with metric radius (Rad-300 m) analysis of Gulangyu from the end of the 19th century to 2013. high tourist numbers) has led to a limited capacity of tourist the National holiday Day in China with a one-hour interval reception on the island. (Fig. 9). The heat maps discussed below reflect the spatial and temporal distribution of tourists and residents produced through 3.2. Verifying the tourists preference via space syntax cell phone use. Both cell phone GPS and location based services are used for providing relatively high resolution behavior re- 3.2.1. Tourists preference from observed big data cords. Considering the fact that the number of tourists heavily The space syntax approach has been proven to be an accurate exceeds the number of residents during the National holiday reflection of morphological structure in the case of Gulangyu. Day, it is reasonable to affirm that the heat maps generally reflect Nevertheless, how well this quantitative description of the distribution of tourists over that day. Outwardly, three morphological structure helps to understanding tourist prefer- important areas, the tourist service area, the Longtou area and ences needs to be discussed further. The observation of real-time the Neicuoao area tend to show high numbers of tourist visiting behaviors through big data mining are collected from Baidu, INC., patterns. one of the largest IT servers in China; the data assembled here Importantly then, this data illustrates that people's behavior from Baidu INC, refers to information collected during one day of inside the city is complex and is affected by numerous factors, Y. Li et al. / Tourism Management 52 (2016) 30e43 37

Fig. 7. This figure demonstrates the overlapping of traditional morphological analysis and space syntax analysis. Left: the overlapping of space syntax analysis at the global level and tourist related function areas, Right: the overlapping of space syntax analysis at the local level within the historic areas. including: morphological structure and minor factors such as at- 3.2.2. Discussion II: space syntax analysis results could reflect mosphere, weather and gate policy, etc. If we only use behavior tourist preferences data for a short period, the analysis is easily disturbed by minor or As stated, overlaying the Baidu heat map for a 12 h period (over side factors. In turn, the overlapping of a long period of data the course of a day) provides a combined result reflecting the collection (with this shorter period) might provide a general indi- spatial distribution of tourists, (as shown in Fig. 10A). In order to cator of behavioral patterns that reflect (to a certain extent) the provide a comparable basis, both the heat map and the space morphological structure discussed in this research. syntax integration are converted to grids as 50*50m and are rep- This feature also explains why we only use tourist behaviors as a resented as an overlay in Fig. 10B. From this analysis, the street verification of space syntax rather than as a foundation for evalu- network integration values as linear data located on the streets can ating tourist organization patterns. Because the evaluation of be compared and analyzed with other polygon data represented by tourist organization requires static data which is inherently related tourist sites, etc. Noticeably then a certain similarity exists between to the morphological structure, space syntax works well in iden- the two results. Pearson correlation further indicates that a high tifying these features compared with a traditional analysis of positive correlation coefficient (0.72) exists between tourist pref- tourist behaviors that are often affected by many side factors. erences and street network integration. In other words, when the 38 Y. Li et al. / Tourism Management 52 (2016) 30e43

Fig. 8. Comparison of integration value among three typical regions.

value of street network integration is high, a high degree of tourist tourist flow will be unlikely to aggregate at these sites. These kinds preferences can be observed. of morphological features will bring benefits to the protection of There are also several unusual grids obtaining high global historic heritage. integration but low tourist preference values (Fig. 10C). On-site Nevertheless, there exist some historic sites that hold higher visits to these unusual grids found that the lack of tourists at numerical values at both the global and local level, such as 1e4: these sites is mainly due to the gated protection policy discussed Former China and South Sea Bank Ltd., Former Southern Fujian above. It is the gated doors that present obstacles for tourists' Christian Bookstore, Former British Asiatic Petroleum Company and visiting and staying in these areas. Therefore, it is reasonable to the Huang Family's Villa. This situation suggests that these places claim that in historic towns like Gulangyu, space syntax results can have the potential to become ‘hot’ tourist sites which will mean in also reasonably reflect tourist preferences to a reasonable degree of turn that large amounts of tourists will probably amass at these certainty. places. The overloading of people at these sites will bring negative effects in terms of historic heritage protection. Therefore, we pro- 3.3. Evaluating tourist organization via space syntax pose that two measures should be taken simultaneously. Firstly, we suggest that there should be some overall adjustment of the layout 3.3.1. Present planned historic sites of urban streets and urban functions so as to avoid the negative The protection of historic heritage, as an important part of his- impacts of mass tourism (large numbers of tourists) at the heritage toric culture protection, will be used as an example in the sections sites. For example, the measure mentioned in the planning policy that follow. According to the 2011conservation management plan we have discussed, contends that the Sanitation Wharf and the of Gulangyu's cultural heritage, there are 53 core elements of cul- Neicuo'ao Wharf to the northwest of the island will also be tural heritage on the island. The distribution of these planned conducive to the reduction of tourist flow and should create a historic sites is shown below in Fig. 11. balance between the east and west sides of the island. By enhancing tourist development at these sites, growing numbers of tourists can 3.3.2. Discussion III: evaluation and suggestions be accommodated. Meanwhile, the Former American Consulate (7) The integration value for each historic site is shown in Fig. 12. and the Xilin Villa (8) which obtains quite low rankings in terms of The dots marked in black represent the planned historic site. Most global integration could be supported by the construction of more of the spaces of the historic sites get medium rankings (referring to road connections. Such measures considered, would take the integration on both the global and local level), which suggest that pressure off other historic sites. A series of suggestions for historic Y. Li et al. / Tourism Management 52 (2016) 30e43 39

Fig. 9. The figure demonstrates the Baidu heat map reflecting behavior intensity between 8:00e19:00 on October 4, 2014 (Baidu Tourism, 2014). heritage protection have been given below in Table 1. building a connection between urban tourism and urban space morphology is still quite rare in contemporary investigations. In 4. Strengths and limitations this paper we have sought to address this gap by taking up Gospodini's (2001) call for a combined consideration of urban 4.1. Strengths design, urban space morphology and urban tourism. In terms of the empirical strengths of our investigation the research case This research contains both methodological and empirical study, Gulangyu, can be regarded as a very good example of many strengths. In terms of the methodological strengths, this research Chinese historic towns that are seeking to develop balances be- attempts to develop the study of space syntax analysis to guide tween sustainable tourist development and historic urban con- sustainable tourist developments. Space syntax, as a strong servation. This research contributes not only to the conservation morphological analysis tool, has been used in the general literature issues of Gulangyu but also a large number of historic towns in to explain a set of morphological-related issues, such as pedestrian China that are confronting similar development-conservation movements and the distribution of small businesses. To date, problems. 40 Y. Li et al. / Tourism Management 52 (2016) 30e43

Fig. 10. The comparison between overlapped Baidu heat map and space syntax.

4.2. Limitations 5. Conclusions

Both methodological and empirical limitations exist in this This study adds to the existing literature by exploring and research, which deserve our attention here. In terms of the meth- applying space syntax analysis and develops research on tourist odological limitations, space syntax analysis of street networks space organization and historic urban conservation. The hy- does not involve the traditional morphological analysis of urban potheses in this paper have been discussed through a set of fabric and functions. As a result an understanding of urban fabric, multi-analysis. Ostensibly it would appear that space syntax functions and their evolution can only be used as a background tool, analysis can provide us with an accurate illustration of the gen- given that space syntax analysis is not well integrated within an eral spatial structure of historic towns. Indeed, as we have seen, approach that analyses the spatial configuration of street networks. tourist preferences on the site of Gulangyu Island demonstrate a A more comprehensive and quantitative approach to this method high correlation in relation to street network integration. Based might provide more rigorous suggestions for the disciplines of on these understandings, several suggestions have been proposed tourist management and historic heritage protection. In terms of to evaluate and guide better tourist space organization and his- the empirical limitations of this research, more attempts at inte- toric heritage protection, through a consideration of the street grating urban morphological understanding and concepts into network integration of tourist sites, tourist routes and historic various tourism planning processes need to be developed. Present sites. applications are still relatively limited and in the future we shall try To conclude, the findings and tests provide insights into the to redress these research imbalances. practices of historic heritage protection and tourist space Y. Li et al. / Tourism Management 52 (2016) 30e43 41

Fig. 11. The location of 53 planned historic sites of Gulangyu. organization; moreover, the research also offers suggestions for moment, work and investigations in this area are still in their infancy planners and investigators who are interested in promoting sus- and there is still a very long way to go before we can understand the tainable development in historic Chinese towns. In the contemporary nature and interactions between space and users of space.

Fig. 12. Spatial integration features of 53 planned historic sites at Gulangyu. 42 Y. Li et al. / Tourism Management 52 (2016) 30e43

Table 1 Suggestions for planned historic sites.

Name Present situation Suggestions

1 Former China and Used as tourist center, current function being fit because of high Enhancing the protection of the historic buildings South Sea Bank Ltd. global integration which means attraction to big tourists flow 2 Former Southern Fujian Used as business occupancy Raising the commercial level so as to restrict tourist flow Christian Bookstore 3 Other Huang Family's Used for residence, buildings damaged seriously Conservative repairing and replacing the function to avoid large amount Villa of tourists and reflect the historic value 4 Former British Asiatic In maintenance Replacing the function to avoid large amount of tourists and reflect the Petroleum Company historic value 5 Sea & Sky Palace Villa It is also a hot tourism site used as an exhibition hall Maintain current situation 6 Former Sanqiutian Jetty Being vacant Introducing new functions that can reflect the historic value 7 Former American Used as a hotel, seriously damaged Conservatively repairing and strengthening connections between the Consulate Former American Consulate and other sites to a proper extent 8 Xilin Villa Used as an exhibition hall Strengthening connections between the Former American Consulate and other sites to a proper extent

Nevertheless, to acquire more guides, future explorations must focus configuration and attraction in urban pedestrian movement. Environment and e on integrating newly-raised techniques to uncover how, in real-time, Planning B: Planning and Design, 20(1), 29 66. Hong, B. (1999). The old time and new light of Xiamen (in Chinese). Xiamen: Xiamen people interact with the urban form. From a big data perspective, GPS University Press. tracking and cell phone data mining of visitors and local residents Ismail, H. N., Chau, L. W., & Rahman, J. (2010). Configuration or attractors? a space fl may provide more understanding and approaches, so as to analyze syntactic inquiry into tourists' movement ow pattern. In The Melaka world heritage site. Malaysia: Centre for Innovative Planning and Development. CIPD tourist space organization. Monograph no. 7. Karimi, K. (2000). Urban conservation and spatial transformation: preserving the ‘ ’ e Acknowledgments fragments or maintaining the spatial spirit . Urban Design International, 5(3 4), 3e4. Karimi, K. (2012). A configurational approach to analytical urban design:‘Space The authors acknowledge the financial support of the Natural syntax’methodology. Urban Design International, 17(4), 297e318. e Science Foundation (NO. 41371388), Fujian Natural Science Foun- Klarqvist, B. (1993). A space syntax glossary. Nordisk Arkitekturforskning, 2,11 12. Lin, Z. (2010). Community development strategy of urbanscenic area: a casestudy of dation (NO. 2015J01226), the Chinese Postdoctoral Science Foun- Kulangsu (in Chinese) City Planning Review, 34(B10), 78e82. dation (NO. 2014M551449), and the Project for Special Funding Li, M., Wu, B., & Cai, L. (2008). Tourism development of World Heritage sites in provided by the Fund for Scientific Research in Colleges and Uni- China: a geographic perspective. Tourism Management, 29(2), 308e319. McKercher, B., Ho, P. S., & du Cros, H. (2005). Relationship between tourism and versities (NO. 2012121033). We are grateful to the Editor and cultural heritage management: evidence from Hong Kong. Tourism Manage- anonymous referees for their valuable comments and guidance. ment, 26(4), 539e548. Peponis, J., Ross, C., & Rashid, M. (1997). The structure of urban space, movement and co-presence: the case of Atlanta. Geoforum, 28(3), 341e358. References Ratti, C. (2004). Urban texture and space syntax: some inconsistencies. Environment and Planning B: Planning and Design, 31(4), 487e499. Administrative Committee of Gulangyu. (2014). Tourist statistics of Gulangyu scenic Ratti, C., Williams, S., Frenchman, D., & Pulselli, R. (2006). Mobile landscapes: using area in May Day 2014. Administrative Committee Report, 4 May http://www.gly. location data from cell phones for urban analysis. Environment and Planning B: cn/zhfxxgk/zhfxxgkml/ (accessed 05.05.14). Planning and Design, 33(5), 727. Baidu Tourism. (2014). Scenic spots of Gulangyu. Baidu Tourism Ranking, 24 May Shoval, N. (2008). Tracking technologies and urban analysis. Cities, 25(1), 21e28. http://lvyou.baidu.com/gulangyu/jingdian/ (accessed 24.05.14). Soja, E. W. (1980). The socio-spatial dialectic. Annals of the Association of American Gospodini, A. (2001). Urban design, urban space morphology, urban tourism: an Geographers, 70(2), 207e225. emerging new paradigm concerning their relationship. European Planning Ståhle, A., Marcus, L., & Karlstrom,€ A. (2005). Place syntax: geographic accessibility Studies, 9(7), 925e934. with axial lines in GIS. In Proceedings of the fifth international space syntax Griffiths, S., Jones, C. E., Vaughan, L., & Haklay, M. (2010). The persistence of sub- symposium; Delft (pp. 131e144). urban centres in Greater London: combining Conzenian and space syntax ap- Turner, A. (2003). Analysing the visual dynamics of spatial morphology. Environ- proaches. Urban Morpholgy, 14(2), 85e99. ment and Planning B: Planning and Design, 30(5), 657e676. Griffiths, S., Vaughan, L., Haklay, M. M., & Emma Jones, C. (2008). The sustainable Turner, A. (2007). From axial to road-centre lines: a new representation for space suburban high street: a review of themes and approaches. Geography Compass, syntax and a new model of route choice for transport network analysis. Envi- 2(4), 1155e1188. ronment and Planning B: Planning and Design, 34(3), 539e555. Gulangyu Island. (n.d.), In Wikipedia, Retrvied March 2, 2015, from http://en. Turner, A., Penn, A., & Hillier, B. (2005). An algorithmic definition of the axial map. wikipedia.org/wiki/Gulangyu_Island. Environment and Planning B: Planning and Design, 32(3), 425e444. Hayllar, B., Griffin, T., & Edwards, D. (Eds.). (2008). City spaces-tourist places: Urban Wang, Y., & Bramwell, B. (2012). Heritage protection and tourism development tourism precincts. London; New York: Routledge. priorities in Hangzhou, China: a political economy and governance perspective. He, Q. (2007). A comparative study of the public concessions in modern China (in Tourism Management, 33(4), 988e998. Chinese) The Editorial Department of the Journal of (Arts & Ye, Y., & van Nes, A. (2012). New ways of spatial modelling: making spatial diagnosis Social Sciences Edition), 6(2), 113e120. in combining space syntax, spacematrix and MXI with GIS of new and old Hillier, B. (1996a). Cities as movement economies. Urban Design International, 1(1), towns. In New urban configurations: EAAE/ISUF international conference; October, 41e60. Delft, the Netherlands. Hillier, B. (1996b). Space is the machine: A configurational theory of architecture. Ye, Y., & van Nes, A. (2013). Measuring urban maturation processes in Dutch and Cambridge, MA: Cambridge University Press. Chinese new towns: combining street network configuration with building Hillier, B. (1999a). Centrality as a process: accounting for attraction inequalities in density and degree of land use diversification through GIS. The Journal of Space deformed grids. Urban Design International, 4(3e4), 107e127. Syntax, 4(1), 18e37. Hillier, B. (1999b). The hidden geometry of deformed grids: or, why space syntax Ye, Y., & van Nes, A. (2014a). The spatial flaws of new towns: morphological com- works, when it looks as though it shouldn't. Environment and Planning B: parison between a Chinese new and old town through the application of space Planning and Design, 26(1), 169e191. syntax, space matrix and mixed use index. A/Z ITU Journal of the Faculty of Ar- Hillier, B. (2002). A theory of the city as object: or, how spatial laws mediate the chitecture, 11(2), 191e208. social construction of urban space. Urban Design International, 7(3), 153e179. Ye, Y., & van Nes, A. (2014b). Quantitative tools in urban morphology: combining Hillier, B., & Hanson, J. (1984). The social logic of space. Cambridge, MA: Cambridge space syntax, spacematrix, and mixed-use index in a GIS framework. Urban University Press. Morphology, 18(2), 97e118. Hillier, B., & Iida, S. (2005). Network effects and psychological effects: a theory of Yu, L., & Liu, Y. (2011). Cultures and community in the process of urban develop- urban movement. In Proceedings of the fifth international space syntax sympo- ment and rehabilitation: analysis of Gulangyu model Xiamen (in Chinese) Ur- sium; Delft. ban Planning International, 25(6), 108e112. Hillier, B., Penn, A., Hanson, J., Grajewski, T., & Xu, J. (1993). Natural movement: or, Y. Li et al. / Tourism Management 52 (2016) 30e43 43

Dr Yuan Li ([email protected]) is an associate Pro- Technology, Delft and is a registered urban designer in the fessor at the School of Architecture and Civil Engineering Netherlands. His research focus is on urban morphology, (SACE), Xiamen University. He graduated from Wuhan urban design and quantitative morphological tools. University (China) in 2001 (Bachelor Degree) and received a double Master Degree in Urban planning and Land Administration (UPLA) (ChinaeNetherlands) in 2004. In Dr Wangtu Xu ([email protected]) is an associate Pro- 2007 he was awarded the degree of Ph.D for his work in fessor at the School of Architecture and Civil Engineering the field of Photogrammetry and Remote Sensing at LIES- (SACE), Xiamen University. He has graduated from Beijing MARS, Wuhan University. He finished his Postdoctoral Jiaotong University in 2010. His research focus is on urban research at the University of North Carolina, Chapel Hill transportation, accessibility and traffic analysis. (UNC); currently he is undertaking a second Postdoctoral research project at Tongji University in the field of Behavior and Tourism.

Longzhu Xiao ([email protected]) is an MSc student in the Department of Urban Planning at the School of Ar- chitecture and Civil Engineering at Xiamen University. Her research interests are: urban morphology, urban social issues and urban design. Dr Andy Law ([email protected]) is a sociolo- gist and town planner at the School of Architecture, Planning and Landscape, Newcastle University. He has published papers, book sections on the subjects of built heritage, imagined histories, nostalgia and social identity. With Professor Andrew Ballantyne he is also author of the very well received book Tudoresque: In Pursuit of the Ideal Home (2011). At present he has been awarded a British Academy Grant to study the Hanfu movement and he in- tends to write a book on the idea of the “modern” in Yu Ye ([email protected]) is a PhD candidate in the contemporary China. Dr Law is currently Degree Pro- Department of Urban Planning and Design in the Faculty of gramme Director of the BA (Hons) degree “Architecture Architecture at the University of Hong Kong. He has a and Urban Planning”, in the School of Architecture, Plan- Master degree in Urbanism from the University of ning and Landscape, Newcastle University.