Data Analysis

Many experiments have been conducted in the past to establish a link between tourist travel behavior and their socio-demographic as well as psychosocial features. Some of the studies have been based on specific geographical areas and concentrate on finding the type and nature of popular activities chosen by the visitors while visiting those locations while others have contributed by merely exploring the tourist behavior in general. Chen, Wang and Prebensen (2015) have explored the activity patterns of tourists based on their demographics and travel companions i.e. travelling alone, with partner or with children. The study conducted in Norway focused around 25 different activities for tourists ranging from short trips, tasting foods, skiing, fishing, theme parks, festivals etc. Yeung et al. (2016) in their study, aimed to find out whether the Japanese Tourists differ in their choice preferences, travel behavior & image perception according to their socio-demographic and travel-related factors. The data was collected through self-administered questionnaires which asked tourists input about their preferences of tourist activities, shopping & travel behavior while visiting Hong Kong. The findings of the study recorded differing travel and behavior patterns in Japanese Tourists according to their age, gender, level of education, type and frequency of visit. Kitajima et al. (2012) conducted a study at a hot spring resort in Western Japan to understand the tourist behavioral selections based on their chronological development. This study deployed a new methodology for analyzing tourists’ in situ behavior, cognitive chrono-ethnography, which incorporates the qualitative understanding of the decision-making process. The findings proved that tourists’ activity patterns are dependent on their intrinsic nature and state of mind rather than extrinsic factors. The activities chosen by tourists were dependent on their motivation levels, past experiences and individual preferences. Another research by Dejbakhsh, Arrowsmith and Jackson (2011) investigated the spatial behavior pattern of International Tourists travelling to Melbourne City in Australia. It demonstrated that these behavioral patterns are the cause of cultural differences among these tourists. The results showed marked differences between tourists from varied cultural backgrounds with respect to their spatial behavioral patterns. These patterns included choice of accommodation location, mode of transport to travel, length and direction of movement and choice of activity at a particular location.

These studies have assisted in the understanding of tourists’ behavior in general and further classifying them according to their needs and preferences. The knowledge gained is partial and limited to the locations tourist prefer and the factors responsible for the tourist choice of destination; however, it fails to provide complete information about tourists visit. Another limitation in these above mentioned studies is that the data was sourced mainly from surveys, questionnaires, individual and group interviews, which were conducted with the volunteered visitors. This data could not give very accurate results due to the human error factor – people cannot recall sometimes about the times / locations and most importantly activities carried out – even if they do, there is still a chance of confusion and misinterpretation (Zoltan & McKercher 2015; Bujosa et al. 2015). Also the data source was limited to only a pre-set number of volunteered visitors, thus the findings could not be generalized to the entire tourist population at the destination.

Tourists’ Mobility Patterns:

Considerable efforts have been spent on finding out tourists’ mobility patterns, aiming to identify the tourists’ mode and path of travel. Klahn et al. (2015) researched the mode and extent of travel of visitors to Munich by carrying out a survey at the Domestic and International Airport Terminal, Central Train Station and Central Bus Stop in Munich. Another study by Yun and Park (2014) on the festival visitors in rural areas showed that most tourists visit only the entrance and central spaces, walking along on the same paths, while only a small number of visitors visit the regional commercial area, including the traditional market, main streets in the downtown area, although their tickets allow them to visit these spaces. Orellana et al. (2012); Beeco et al. (2013) and Bauder et al. (2015) observed the effects of travel preparation on tourists mobility patterns in and around the destination and found that the tourists who prepared their visit before travelling to a destination were the ones covering most of the attractions and activities at a particular destination.

Another popular area of research with regards to tourists, is exploring their spatial and temporal patterns when visiting places of interest. Some researchers have confined their research on finding the tourist spatio-temporal patterns while in transit or at a destination (Bauder 2015; Birenboim 2013; Beeco et al. 2013; Long et al. 2013; Orellana et al. 2012). Others have gone deeper and extracted useful information about the tourist trajectories, activity location with respect to time spent at each activity and difference with regards to first time visit and repeat visitation (Bujosa et al. 2015; Cantis et al. 2016; Hallo et al. 2016; Kadar 2014; Maghrebi et al. 2015). Birenboim et al. (2013) used GPS technology to track and record the time-space trajectories of tourists at a theme park in Spain. The research studied the mass activity patterns of tourists during the 24 hours period while visiting the Theme park location. The findings resulted in the time spent at each activity in the park and the activity pattern of the tourists. 5 main categories of activity were Rides, Shows, Restaurants, Shops, Games. The findings also showed the diurnal and intra-diurnal activity patterns. Another research by Cantis et al. (2016) used a segmentation approach based on movement patterns of cruise passengers in the city of Palermo. The results of the study can be very useful to the cruise line companies as well as destination managers and tour operators in helping them devise routes and plan activities to gain maximum satisfactory feedback from their tourists and at the same time maximize profits for themselves. Similarly Orellana et al. (2012) used the GPS tracking data in a National Park in Netherlands to study the collective movement patterns of tourists. The focus was on two kinds of movement patterns – MSPs (Movement Suspension Patterns) and GSPs (Generalized Sequential Patterns). The MSPs was used to detect the main places visited by people in a recreational park while the GSPs was used to establish the sequence in which these places were visited. Grinberger, Shoval and McKercher (2014) made another interesting addition to the literature by presenting a conceptual framework to describe and understand tourists’ spatio-temporal behavior, according to which the effective reach of an individual is defined by time–space constraints and the path taken by the individual. It is evident that the research under these topics has been conducted to draw out useful and real-time information from tourists by collecting data through their hand-held GPS devices. The GPS/GIS device coverage and signals can pose a limitation to the collected data. The GPS device may not be accessible in many remote areas and there is a high chance of disruption in the signals’ quality and strength (Orellana et al. 2012; Hallo et al. 2016). Furthermore, the past work using this GPS technology was unable to reveal the actual activities undertaken by tourists at each visited location.

Tourist Activity Analysis:

There have been some attempts made in the past in studying the touristic flows and predicting tourist future destinations. One of the early contributions done in this respect was from Yang, Fik and Zhang (2013) who focused on the decision making process of tourists and tried to model their next destination by using the nested logit model. This model assumed the ‘utility maximization’ as the tourists demand and targeted on only one subsequent destination, unlike tourists, who can visit multiple subsequent destinations. Another research was done by Zheng, Huang and Li (2017) who tried to model the tourist next destination through a survey group who collected data on tourists’ intra-attraction spatial-temporal behavior and demographic characteristics using handheld GPS tracking devices and activity diary questionnaires. One of the latest research done in this category is by Vu et al. (2015) who explored the travel behaviors of tourists in Hong Kong by using the data from geo-tagged photos uploaded on Flicker. The tourists’ movement trajectories were highlighted and patterns were drawn to indicate the most popular tourists’ destinations. The location preferences was also categorized with respect to 2 main groups Asian n Westerns. A similar study was carried out while exploring visitors’ activities in Hong Kong Parks (Vu et al. 2016) and Temples (Leung et al. 2016). The Twitter ‘s API has been used by Chua et al. (2016) in which the geo-tagged social media data is used to categorize tourists flow in Italy. This research used the geo-tagged social media data from Twitter to characterize spatial, temporal and demographic features of tourists’ flow in Cilento, Southern Italy. The study focused on 3 main areas which are: the tourists’ profiles, tourists’ travel patterns in the region, tourists’ attraction in the region and their popularity. The geo-tagged photos on Flickr have been used in many other studies (Kadar 2014). However, Flickr only gives the geo-coordinates of the photos without giving any information about the actual place, its type, category and user comments associated with it. What the tourist actually did at that particular location, which activities they were involved in, and how much time they spent at that particular activity cannot be assessed via the data obtained from Flickr.

Summary:

A good understanding of tourists’ activities is important for tourism management to ensure they provide the best service, meet tourist expectation and also gain repeat visitation. The past studies have been concentrated around tourists’ decision making process, travel trajectories and movement patterns. The knowledge gained from the past research has been beneficial in understanding tourists’ behavior, however, little information is gained about tourist activity at a particular destination which is very important for tourism management and can assist in many ways. This research will fill the gap in the tourism literature by studying tourist activities in the sequential order, thereby providing rich information about tourist choices, preferences and decisions while visiting a particular destination.

METHODOLOGY AND CONCEPTUAL FRAMEWORK

This section deals with the experiment and the analysis of the study, for the purpose of which the Knowledge Discovery in Databases (KDD) framework has been applied for efficient and timely evaluation of data. One of the early contributors to this framework, Osama Fayyad (1996), has described ‘Knowledge Discovery’ as the non-trivial process of identifying valid, novel, potentially useful, and ultimately understandable patterns in a large dataset. The main idea in KDD is to discover a high level knowledge (abstract knowledge) from lower levels of relatively raw data, or to discover a higher level of interpretation and abstraction than those previously known. Since our dataset contains a high volume of textual information, we chose to apply KDD for the process of knowledge discovery. The major steps in the KDD process are as follows:

1. Data Collection: Creating the main data set for the purpose of the KDD.

2. Data Pre-processing: Cleaning and Data processing.

3. Data Mining or Data Analysis: Applying the data-mining algorithms such as clustering, classification, regression, in ways which match with the original goal of the KDD process. Data mining and searching for interested patterns in the data.

4. Evaluation of Results: Interpretation of the result or found pattern. A good clustering or classifying approach should explain the result of such an approach. Prepare a new set of knowledge for future analysis, utilization or discovery purposes.

Fig.1 is a representation of the KDD process for this research which highlights the key areas and the main steps involved in it.

Fig 1. Proposed Work Flow Chart

1. Data Collection:

We will collect the Foursquare Check-in Data through Twitter Streaming. Foursquare is one of the largest location based social network (LBSN), having approximately 3 million check-ins per day. It not only provides information about the venue type and location but its game-like structure encourages its users to continue check-ins repeatedly. To accomplish this, the Twitter’s Streaming Application Programming Interface (API) will be used to collect the publicly available tweets with pre-set latitude and longitude coordinates. Each check-in has the following attributes: check-in ID, user ID, time and geographic coordinate (latitude and longitude), category and subcategory of the check-in’s location, i.e., the type of place where it occurred.

The dataset will comprise of the most frequent users of Foursquare / Twitter, the ones who like to record their activity on a regular basis – at least 3-5 times per day. A method will be designed to extract the check-ins for some specific users using the Twitter and Foursquare API. This data will be collected over the sample period of atleast 4-6 months. The dataset will be extensively drawn from the posts of the tourists who use Foursquare on a regular basis to record their everyday activity. This method of data collection provides us many additional characteristics of tourist transition pattern which were not included in previous studies.

2. Data Pre-processing:

The raw data obtained from the Twitter’s API will be processed to make sure there is no noise, outliers, inconsistent and incomplete entries into the final database. At first the data will be run through the data cleaning routines which will smooth the noise, identify and remove outliers and resolve any inconsistency and double entries in the database. The cleaned data will then be transformed into easy to read and interpret Excel files with fields representing various attributes of the data such as the Twitter ID, Time and location of Tweet, Text, Language, Favorites and any hash tags associated with the tweet. However, due to the limited nature of the research, only the data revealing information about the spatio-temporal attributes of tourists will be scrutinized for e.g. the user profile can be used to make the distinction between several tourists’ traits like age groups, gender and nationality; the location category data can specify the activity patterns of the tourists; and the time can classify the temporal patterns related to events that occur at certain times.

3. Data Mining Process:

A complete understanding of tourist activity analysis requires knowledge of the activity recognition followed by the activity pattern discovery. In this research, the first step will be to accurately detect the tourists’ activities from the data set extracted from Twitter Live Steaming. This will be accomplished through customized program, which will extract the venue category of each tweet belonging to the foursquare check-ins. These venue categories are the depicters of exact activity that the tourist was involved in. This information will lead to further pattern discovery and a source of data mining process. Next the source files will be created for the selected algorithms which will reveal some interesting patterns and knowledge discovery. Thus, the selected data mining techniques are discussed in the following paragraphs:

3.1 Association Rule Mining:

3.2 Sequential Pattern Mining:

The sequential pattern mining is a data mining task specialized for analyzing sequential data, to discover sequential patterns. More precisely, it consists of discovering interesting subsequences in a set of sequences, where the interestingness of a subsequence can be measured in terms of various criteria such as its occurrence frequency, length, and profit. For the purpose of mining sequential patterns, two algorithms have been chosen namely SPAM and LAPIN. The SPAM algorithm out performs SPADE & PrefixSpan by finding all frequent sequences from a list of transactions in a fast and efficient manner. The data represented in a bitmap form results in efficient counting at each recursive step of the sequence. Another algorithm that has been proposed for this study is LAPIN (LAst Position INduction) sequential pattern mining, which unlike SPAM uses the last position of an item s to judge whether the frequent sequence pattern could be appended with the item s to make it a frequent length pattern. This can be beneficial for this study since we aim to find the frequent activity patterns of tourists and could be further researched into finding the most popular future tourists activity patterns. Another key benefit of using LAPIN is its fast turnover, since it searches only a small portion of the database by recording the last position of each item in each sequence. These can help to cluster the venues and also address the most popular activities at a location based on the number of check-ins and crowd level. This technique has been used by Long et al. (2013) who focused their study to understand human travel behavior through LSBNs and were successful in finding the diversity of user-behavior when they travel to a destination.

4. Evaluation of Results:

It is assumed that the study will reveal various interesting patterns related to tourist activities, however due to the confined nature of this research we will extract and interpret only the sequential activity patterns of tourists at diverse locations. It is evident from previous research that tourist behavior can be volatile and easily affected by a series of inter-twined factors (Bauder & Freytag 2015; Beeco et al. 2013; Chen et al. 2015; Dejbakhsh et al. 2011; Klahn et al. 2015). Thus, we attempt to find some regularity in the extracted patterns. Thereby contributing to the knowledge about tourist behavior related to the particular order of activities they prefer to choose. This might also be dependent on some extrinsic as well as intrinsic factors. For example tourist visiting a place might choose to take part in recreational activities like sight-seeing, and afterwards go for dining and after that choose to shop around the place. Studying these series of activities carried out by tourists on a large scale can give us an insight into tourist trends, choices and preferences which can assist in the understanding of tourist behavior. The knowledge gained can be applied for the strategic development of tour management and can lead to more sophisticated tour packages and travel itineraries.

To study tourist activity patterns, we need to determine how close or similar a tourist’s activity pattern is relative to that of another tourist. An effective measure of the distance or similarity between the tourist activities should take into account many characteristics of tourist activities apart from spatial and temporal dimensions. Research has proven that tourists conduct different activities at certain times, therefore, when comparing individual activity patterns, differences in the attributes of these activities (e.g., type and purpose), should also be captured (Helal, Kim & Cook 2010). The interdependency among these dimensions needs to be maintained in a distance or similarity measure (e.g., certain activities can take place only at certain places and/or at certain times). When comparing activity patterns, the distance measure should also be able to compare structural differences in tourist activities and their contextual variables (e.g., certain activities have to be performed before specific other activities) (Helal, Kim & Cook 2010). Thus, the tourist activity patterns will unfold over time.

ETHICS APPROVAL

In this study the data is mainly collected from the geo-tagged tweets in twitter. Since, all the tweets used in this study are publicly available, there are no privacy concerns or issues to be addressed.

CONCLUSION

This research is based on the application of social media (Twitter / Foursquare) data, using data mining and clustering techniques, to extract the sequential activities of tourists at different destinations. The data, collected over the sample period, will be trimmed, cleaned, processed and clustered into the main categories of tourists’ activities as revealed from the information collected through social media. Unlike similar studies in this area, this study focuses on the analysis of data gathered through twitter’s API rather than only using geo-tagged data or sentiment analysis. Thereby, it is believed that this study will reveal more up-to-date and comprehensive report on tourist sequential activity patterns which has not been carried out before in the field of tourism.

TIMELINE

For the sake of convenience and referencing the following timeline has been constructed for this research:

Fig 2. Timeline of the research process.

REFERENCE LIST

Bauder, M 2015, ‘Using GPS supported speed analysis to determine spatial visitor behavior’, International Journal of Tourism Research, vol. 17, pp. 337-346.

Bauder, M & Freytag, T 2015, ‘Visitor mobility in the city and the effects of travel preparation’, Tourism Geographies, vol. 17, no. 5, pp. 682-700.

Blyablina, A 2015, ‘The contribution of guides in developing tourist experiences during historical theatrical tours: The case of Stockholm ghost walk’, European Tourism Research Centre – Department of Social Sciences.

Birenboim, A, Clave, SA, Russo, AP & Shoval, N & 2013, ‘Temporal activity patterns of theme park visitors’, An International Journal of Tourism Space, Place and Environment, vol. 15, no. 4, pp. 601-619.

Beeco, JA, Huang, WJ, Hallo, JC, Norman, WC, McGehee, NG, McGee, J & Goetcheus, C 2013, ‘GPS tracking of travel routes of wanderers and planners’, Tourism Geographies, vol. 15, no. 3, pp. 551-573.

Bujosa, A, Riera, A & Pons, PJ 2015, ‘Sun and beach tourism and the importance of intra-destination movements in mature destinations’, Tourism Geographies, vol. 17, no. 5, pp. 780-794.

Cantis, S, Ferrante, M, Kahani, A & Shoval, N 2016, ‘Cruise passengers’ behavior at the destination: Investigation using GPS technology’, Tourism Management, vol. 52, pp. 133-150.

Chen, JS, Wang, W & Prebensen, NK 2015, ‘Travel companions and activity preferences of nature-based tourists’, Tourism Review, vol. 71, no. 1, pp. 45-56.

Chua, A, Servillo, L, Marcheggiani, E & Moere, AV 2016, ‘Mapping Cilento: using geotagged social media data to characterize tourist flows in southern Italy’, Tourism Management, vol. 57, pp. 295-310.

Dejbakhsh, S, Arrowsmith, C & Jackson, M 2011, ‘Cultural influence on spatial behavior’, Tourism Geographies, vol. 13, no. 1, pp. 91-111.

Edwards, D & Griffin, T 2013, ‘Understanding tourists’ spatial behavior: GPS tracking as an aid to sustainable destination management’, Journal of Sustainable Tourism, vol. 21, no. 4, pp. 580-595.

Grinberger, AY, Shoval, N & McKercher, B 2014, ‘Typologies of tourists’ time-space consumption: a new approach using GPS data an d GIS tools’, Tourism Geographies, vol. 16, no. 1, pp. 105-123.

Hallo, JC, Beeco, JA, Goetcheus, C, McGee J, McGehee, NG & Norman, WC 2016, ‘GPS as a method for assessing spatial and temporal use distributions of nature-based tourists’, Journal of Travel Research, vol. 51, no. 5, pp. 591-606.

Helal, S, Kim, E & Cook, D 2010, ‘Human Activity Recognition and Pattern Discovery’, IEEE Pervasive Computing, vol. 9, pp. 48-53.

Kadar, B 2014, ‘Measuring tourist activities in cities using geotagged photography’, Tourism Geographies, vol. 16, no. 1, pp. 88-104.

Kitajima, M, Tahira, H, Takahashi, S & Midorikawa, T 2012, ‘Understanding tourists’ in situ behavior: A cognitive chrono-ethnography study of visitors to a hot spring resort’, Journal of Quality Assurance in Hospitality & Tourism, vol. 13, pp. 247-270.

Klahn, DT, Roosen, J, Gerike, R & Hall, CM 2015, ‘Factors affecting tourists’ public transport use and areas visited at destinations’, An International Journal of Tourism Space, Place and Environment, vol. 17, no. 5, pp. 738-757.

Leung, R, Vu, HQ, Rong, J & Miao, Y 2016, ‘Tourists visit and photo sharing behavior analysis: A case study of Hong Kong temples’, International Tourism and Hospitality, pp. 197-209.

Li, Z, Deng, S & Moutinho, L 2014, ‘The impact of experience activities on tourist impulse buying: an empirical study in China’, Asia Pacific Journal of Tourism Research, vol. 20, no. 2, pp. 191-209.

Li, Y, Xiao, L, Ye, Y, Xu, W & Law, A 2016, ‘Understanding tourist space at a historic site through space syntax analysis: The case of Gulangyu, China’, Tourism Management, vol. 52, pp. 30-43.

Long, X, Jin, L & Joshi, J 2013, ‘Towards understanding travel behavior in Location-Based Social Networks’, Symposium on Selected Areas in Communications, School of Information Science.

Maghrebi, M, Abbasi, A, Rashidi, TH & Waller, ST 2015, ‘Complementing travel diary surveys with Twitter Data: Application of text mining techniques on activity location, type and time’, IEEE 18th International Conference on Intelligent Transportation Systems, School of Engineering & IT, UNSW.

Majid, A, Chen, L, Mirza, HT, Hussain, I & Chen, G 2015, ‘A system for mining interesting tourist locations and travel sequences from public geo-tagged photos‘, Data and Knowledge Engineering, vol. 95, pp. 66-86.

Mathew, W, Raposo, R & Martins, B 2012, ‘Predicting future locations with Hidden Markov Models’, Instituto Superior Tecnico, Portugal.

Orellana, D, Bregt, AK, Ligtenberg, A & Wachowicz, M 2012, ‘Exploring visitor movement patterns in natural recreational areas’, Tourism Management, vol. 33, pp. 672-682.

Ting, HX & Hu, WB 2012, ‘Intra-attraction tourist spatial-temporal behavior patterns’, Tourism Geographies, vol. 14, no. 4, pp. 625-645.

Vu, HQ, Li, G, Law, R & Ye, BH 2015, ‘Exploring the travel behaviors of inbound tourists to Hong Kong using geotagged photos’, Tourism Management, vol. 46, pp. 222-232.

Vu, HQ, Leung, R, Rong, J & Miao, U 2016, ‘Exploring Park Visitors’ Activities in Hong Kong using geotagged photos’, International Tourism and Hospitality, pp. 183-196.

Versichele, M, Groote, L, Bouuaert, MC, Neutens, T, Moerman, I & Weghe, NV 2014, ‘Pattern mining in tourist attraction visits through association rule learning on Bluetooth tracking data: A case study of Ghent, Belgium’, Tourism Management, vol. 44, pp. 67-81.

Wang, D, Qian, J, Chen, T, Zhao, M & Zhang, Y 2014, ‘Influence of the high-speed rail on the spatial pattern of regional tourism – taken Beijing-Shanghai high-speed rail of China as example’, Asia Pacific Journal of Tourism Research, vol. 19, no. 8, pp. 890-912.

Yeung, MW, Kim, S & Schuckert, M 2016, ‘Japanese tourists to Hong Kong: their preferences, behavior, and image perception’, Journal of Travel & Tourism Marketing, vol. 33, pp. 730-741.

Yang, Y, Fik, T & Zhang, J 2013, ‘Modeling sequential tourist flows: Where is the next destination?’, Annals of Tourism Research, vol. 43, pp. 297-320.

Yang, Y & Wong, KKF 2012, ‘Spatial distribution of tourist flows to China’s cities’, Tourism Geographies, vol. 15, no. 2, pp. 338-363.

Yun, HJ & Park, MH 2014, ‘Time-space movement of festival visitors in rural areas using a smart phone application’, Asia Pacific Journal of Tourism Research, vol. 20, no. 11, pp. 1246-1265.

Zoltan, J & McKercher, B 2015, ‘Analysing intra-destination movements and activity participation of tourists through destination card consumption’, Tourism Geographies, vol. 17, no. 1, pp. 19-35.

Zheng, W, Huang, X & Li, Y 2017, ‘Understanding the tourist mobility using GPS: where is the next place?’, Tourism Management, vol. 59, pp. 267-280.

Zhu, YX & Qun, W 2014, ‘Exploratory space-time analysis of inbound tourism flows to china cities’, International Journal of Tourism Research, vol. 16, pp. 303-312.

Zheng, Y & Zhou, X 2011, ‘Location-based social networks: Users. In Computing with Spatial Trajectories’, Report Springer, viewed on 20th Dec 2016.

Zheng, Y 2016, Mobile LBS Market in the US 2015-2019 – Report Linker Review.

pg. 13