=Paper=
{{Paper
|id=Vol-2620/paper1
|storemode=property
|title=A Mobility Data Model for Web-Based Tourists Tracking
|pdfUrl=https://ceur-ws.org/Vol-2620/paper1.pdf
|volume=Vol-2620
|authors=Thouraya Sakouhi,Jamal Malki,Jalel Akaichi
|dblpUrl=https://dblp.org/rec/conf/balt/SakouhiMA20
}}
==A Mobility Data Model for Web-Based Tourists Tracking==
https://ceur-ws.org/Vol-2620/paper1.pdf
A Mobility Data Model for Web-Based Tourists
Tracking
Thouraya Sakouhi1 , Jamal Malki2 , and Jalel Akaichi3
1
Université de Tunis, ISG, LR99ES04 BESTMOD, 2000, Le Bardo, Tunisia
thouraya.sakouhi@gmail.com
2
Université de La Rochelle, L3i Lab., 17000, La Rochelle, France
jmalki@univ-lr.fr
3
Bisha University, College of Computer Science, 255, Bisha 67714, Saudi Arabia
jakaichi@ub.edu.sa
Abstract. Tracking tourists activities at different levels of their jour-
neys provides an overview on their mobility and a comprehension of their
behavior and preferences. Most information related to tourism services
and tourists are collected and stored through web platforms. In fact, self-
drive tourists access touristic information available on the web to plan for
their trips. Accordingly, tourism professionals track their requirements
in touristic information and then their mobility. Yet, since touristic in-
formation is managed at a territorial level, tracking tourists’ movement
by tourism professionals, out of their territory, is not a straightforward
task. Accordingly, the latters do not have a complete overview of tourists
movements. Throughout this paper authors will start by discussing mo-
bility data capture through the web and the related challenges. Then,
they’ll introduce an integrated mobility data model for tracking tourists.
Keywords: Mobility data, Trajectory, Semantic modeling, Ontology,
Web-based tracking.
1 Introduction
Nowadays, wireless sensor devices such as GPS, RFID and mobile phones are tied
to mobile entities to trace their movements and generate what is called mobility
data. Once analyzed, this data is expected to express a lot of semantics about
mobile entities’ activities and their dynamics. For instance in the case of tourists
mobility data, this motivated large number of research works to investigate novel
approaches for extracting from tourists mobility data activities performed during
their travel. However, tourists mobility data generated in this way is not always
available for researchers and this is due to privacy issues.
From another side, public institutions detain information about touristic offer
and tourists demand and possess efficient tools for the managements of related
PLAIBDE Project. Programme Opérationnel FEDER-FSE de la région Nouvelle
Aquitaine - France
Copyright © 2020 for this paper by its authors. Use permitted under Creative
Commons License Attribution 4.0 International (CC BY 4.0)
1
data. These tools represent the tourism ecosystem. Such data encompasses infor-
mation about touristic offer obviously but also data related to the tourist him-
self, and his requests according to the touristic offer. It then presents a valuable
source of information about tourists mobility during their journey and present
an alternative of the mobility data captured by positioning devices. Yet, due to
the information sparsity within and across different destinations, tourists do not
have a complete and clear overview of touristic offer neither public institutions
and service providers on tourists preferences. Then integration efforts have to be
made to unify touristic information related to destinations. State of the art works
consider only the integration of touristic offer data in their proposed models for
tourism field. In the remainder of this paper, we’ll present our proposed model
of tourist mobility data. This model will give a complete view of the tourism
field data. We’ll make use of the DATAtourisme ontology being a new platform
for integrating touristic information allover France.
Accordingly, this paper is organized as follows. In the second section, authors
will define mobility data capture and the related challenges. In the third section,
authors will present the tourism ecosystem and the data flow within. After that,
authors will overview state of the art touristic data models and the DATA-
tourisme ontology. Then, we will introduce our tourist mobility data model.
Finally, we will discuss challenges raised by the introduction of our model.
2 Mobility Data Capture
2.1 Sensor-Based Mobility Data Capture
Nowadays, sensor devices are integrated to all kinds of moving entities: vehicles,
animals and humans. However, while animals and vehicles’ mobility data are
readily available and accessible, this is not the case for humans mobility data,
which is considered as sensitive by the law. Indeed, personal information are
protected by the means of a set of norms and laws in different countries. For
example, in French law, the CNIL (Commission Nationale de l’Informatique et
des Libertés), a committee responsible for protecting personal data in France,
and RGPD (Réglement Génénral sur la Protection de Données) for the European
Union, impose regulations on expanders of personal information, thus limiting
the dissemination of this data, its use and its processing to avoid violation of
individuals’ privacy. Consequently, this legal constraints made access to this type
of data complicated despite its abundance. From the other hand, mostly data
collection for research made in the context of human mobility data modeling and
analysis requires the recruitment of participants, volunteering or payed, to wear
positioning sensors as they go about day-to-day activities. So even the collection
of mobility data is not a straightforward task.
2.2 Web-Based Mobility Data Capture
By accessing websites and using mobile applications, users transmit indirectly
different types of their data to first and third parties. The data transmission
2
observed between clients and servers helps parties to keep state of the user over
time. Such data is historized and associated with the website visitor to be rec-
ognized later. Among the captured data, eventually data about the location
of the user is collected, then his mobility according to space and time. Indeed,
web/mobile applications’ managers track their users online using different meth-
ods including: cookies, HTTP metadata, device fingerprinting, local storage, etc.
Originally, online behavioral data collection is limited to connecting users across
multiple websites on one device. However, today companies are finding ways to
correlate user’s behavior, by third parties, across different devices by identifying
linked devices a unique individual uses. This tracking method is referred to as
cross-device tracking [2].
3 Mobility Data Within the Tourism Ecosystem
As tourism and travel are known to be information intensive domains, travel-
related information is available abundantly in the web [3]. The interaction be-
tween tourists and tourism professionals generates information exchange be-
tween the two stakeholders. While tourists ask for information about their jour-
neys, tourism professionals ask for information about the tourist, his location,
where he intends to go and more, to communicate him later information he
requested. Website owners register users information [2] for analytical and ad-
vertising purposes. These users’ informations constitute an alternative tool for
tracking tourists movement implicitly. The management of territorial touristic
information requires a set of tools including, mainly the Touristic Information
System (TIS), Customer Relationship Management (CRM) and Reservation Sys-
tem (RS), which represent the tourism ecosystem illustrated in figure 1.
Fig. 1. Tourism ecosystem: captured data flow and processing
3
In France, touristic information is organized by geographical department.
Actually, departmental TIS is fed by the DCT (Departemental Commitee of
Tourism) [8] and its partners, the former is then the socle of all DCT publica-
tions. This pooling of resources allows a completeness of the offer and a saving
of time. However, a lot of touristic service providers at the territory level still do
not use the official DCT portal to share their touristic offers. Accordingly, the
diversity and variety of information sources on the web related to one territory,
evenmore all over France, makes a huge amount of information. And as quantity
of available information is large, self-drive tourists pose a big challenge in effi-
ciently accessing and managing the tourism related data and find out about their
desired destinations in France. Consequently, the actual tourism ecosystems data
in France has the following properties:
– Proprietary: it is the property of the party that generates them: the depart-
ment, the service provider, etc, and are not accessible to the other sector
stakeholders.
– Heterogeneous: the different TISs do not have the same data structure, nor
the same semantic and syntactic representation.
– Territorial: limited to the representation of tenders at the concerned territory
level.
Accordingly, efforts for the integration of territorial tourism ecosystems data
throughout France are essential. On the one hand this will permit to improve the
visibility online of touristic offer and increase its availability. On the other hand,
this will give a complete view of tourists, their interactions with information
sources and then their mobility within French departments. So, an upgrade to a
unified tourism data model at a national level is highly required. In the sequel,
authors review the more common proposals of tourism data integration models.
4 Tourism Data Modeling: An Overview
In the following, we concentrate in French initiatives of national tourism data
integration. TourInFrance (TIF), TIFSem and DATAtourisme are from of those
research thrusts that proposed solution for touristic data globalization at a na-
tional level in France. We provide a description of these models in the following.
4.1 TourInFrance
TIF is a norm that was initiated in 1999 by the French State Secretariat for
Tourism to facilitate touristic data exchange between different TISs in France [5].
This norm is used by national tourism offices, departmental tourism committees
and tour operators. It is represented by a metadata format describing tourist
information in detail [4]. In 2004, TIF format has evolved to XML technologies
in order to facilitate the uploading of data on websites and the exchange of in-
formation between different systems. However, since 2005 TIF norm has stopped
4
evolving. Accordingly, each of the aforementioned tourism professionals adapted
the touristic content to their special requirements. Furthermore, with the recent
advances of web technologies and its upgrade to the semantic web, TIF became
obsolete and useless.
4.2 TIFSem
TIFSem [7] is a tourism ontology developed in the context of the TourinFlux
project [6] that proposes a set of tools that allow tourism professionals to handle
unified datasets and improve their display on the web. They developed for that
an ontology named TIFSem that is based on the TIF norm, representing a set of
concepts and relations for tourist resources. Besides, other sources are consulted
to collect concepts related to the tourism domain.
4.3 DATAtourisme
DATAtourisme [1], a universal language to describe points of interest for tourists,
was proposed by the French government. The latter intends to create a new
platform for tourism in France in a national level, and this is by integrating data
from various TISs at a territorial level. This permits then to provide touristic
information requirements using Semantic Web Services. In brief, DATAtourisme
is a shared system that allows territories to put their touristic data in Open
Data. DATAtourisme’s system has the following characteristics:
– Semantic standard since the meaning of each element of the vocabulary is
formal and consensual.
– Open and simplified because available to all and easily reusable.
– Universal because generic enough to apply to any type of tourism content
which made possible the harmonization of the vocabulary of each TIS.
– Scalable, this vocabulary can easily evolve to adapt to new needs.
– Language independent and can be used to describe multilingual content.
– Interoperable because it interacts with other authoritative international vo-
cabularies, for example DATAtourisme is natively compatible with Schema.org,
INSEE, DublinCore, FOAF vocabularies.
Accordingly, later on, we will make use of the DATAtourisme ontology for
the construction of our tourist mobility data model.
5 Tourist Mobility Data Model
A tourist is considered as a mobile entity since he moves across different loca-
tions within his destination: museums, hotels, restaurants, etc. Such locations
are considered as Points Of Interest (POI) according to the tourism context.
These POIs represent also the touristic offer exposed in the web or by the way
of the different available channels. While a tourist seeks touristic offer in the
5
web, he interacts with touristic content by sending offer requests, making reser-
vations, etc. Consequently, this interaction, once analyzed in a global context,
provides tourism professionals with valuable information about tourists, their
touristic interests and then their mobility before, within and after visiting their
destinations.
To do so, we start throughout this work by modeling the tourist mobility data
in this context. Our objective is to represent tourist mobility data in a generic,
complete and simple model. We intend to categorize the principle entities repre-
senting data within the tourism ecosystem previously described. Throughout this
section, we introduce the formal model to represent data flow within the tourism
ecosystem between the different stakeholders taking part in the data exchange
mechanism including tourists and tourism professionals. Authors refer to this
model as the Tourist Mobility Data Model (TMDM). Actually, the observation
of this data flow brings out three types of data: data that represents the touris-
tic offer, data about the set of events triggered when the tourist seeks touristic
offer and data that defines the tourist. An in-depth study of the properties of
each of these types of data leads us to the TMDM. Accordingly, the TMDM is
mainly composed of 3 sub-models: touristic data space (TDS), event space (ES)
and tourist space (TS). In the following subsections, we explain separately each
components.
5.1 Touristic Data Space Model
Touristic data constitutes all data related to the touristic offer including: accom-
modation, entertainment activities, transportation, catering, etc. We consider
DATAtourisme since it is then used by almost all governmental and private
tourism institution in the country. The DATAtourisme ontology summarizes the
touristic content in the POI concept as it represents the minimal definition of
any touristic object. POI is defined as any touristic product that deserves to be
described and valued. It is a touristic element which is managed by an agent
and which could be consumed through products and services [1]. This is the
minimal class to instantiate for a product to be managed in the DATAtourisme
information system. A POI concept decomposes into 4 sub-concepts: Product,
Tour, Entertainment and Event and Place of Interest.
5.2 Event Space Model
The event space represents data resulting from the set of interactions between
tourists and touristic data resources made available to them via the different
contact points (web/mobile applications, e-mail, phone, tourism offices, etc).
Commonly, if an individual is using a web/mobile application, it is very likely
that he is being tracked by the application developers. This actually refers to
collecting and passing data about user actions when interacting with the ap-
plication. Each related user action triggers an event. The set of collected data
include data about the user: location, time, age, gender, coordinates, device
6
information, interests, etc, and information about his interaction with the ap-
plication: information sought, time spent on every screen, clicks, ad views, etc.
In the case of tourists interacting with touristic data, events are recorded within
the CRM database. An event in this case represents a request from a tourist and
the response to this request.
5.3 Tourist Space Model
The tourist space represents the set of tourist’s intrinsic data that defines him:
location, identity, coordinates, genre, interests, etc. The tourist, stands for the
moving entity in the context of our application.
The association between the TDS and ES sub-models is performed between
the Request and Response concepts from one side and PointOfInterest concept
from the other side. The association between the ES and TS sub-models is
performed via the Tourist and Event concepts. The TMDM is illustrated in
the figure 2.
Fig. 2. Tourist Mobility Data Model
6 Discussion
The transition from the territorial to the national level after DATAtourisme
will raise, mainly, scientific and technological challenges about the characteri-
zation of the touristic mobility data in the future. Actually, with TIS at the
territorial level, it was not possible to link tourists behaviors across different
platforms/websites they access to seek information or make reservation due to
the variety of stakeholders taking part in the presentation of the touristic offer,
which made difficult identifying uniquely the tourist. With DATAtourisme, since
7
all touristic information will be available and structured in the same vocabulary,
cross-platform tracking of tourists will be affordable, and tourists will be eas-
ily recognizable at a national level. Even more, it will be possible for tourism
professionals to track tourists mobility before, during and after visiting their ter-
ritory. This will provide a more complete view into tourist’s behavior and could
be useful for a set of applications. Thereupon, this raises in turn an economical
challenge about the development of tourism services and touristic offer and de-
mand, leading to the prosperity and development of the tourism sector in the
country, the latter being one of the most important economic activities, but also
to the satisfaction of tourists with regards to the services offered becoming more
targeted and of high quality.
Acknowledgment. This work is carried out thanks to the support of the Eu-
ropean Union through the PLAIBDE project of the FEDER-FSE operational
program for the Nouvelle-Aquitaine region. The project is supported by aYa-
line company, with partners: LIAS-ENSMA laboratory in Poitiers, and the L3i
laboratory at La Rochelle University.
References
1. Ontologie datatourisme v1.0 documentation de référence. Technical report, 01 2017.
https://www.datatourisme.gouv.fr
2. J. Brookman, P. Rouge, A. Alva, and C. Yeung. Cross-device tracking: Measurement
and disclosures. Proceedings on Privacy Enhancing Technologies, 2017(2):133–148,
2017.
3. D. K. Chiu, Y. T. Yueh, H.-f. Leung, and P. C. Hung. Towards ubiquitous tourist
service coordination and process integration: A collaborative travel agent system
architecture with semantic web services. Information Systems Frontiers, 11(3):241–
256, 2009.
4. T. et TERRITOIRES. Ressources documentaires tourinfrance.
5. F. Soualah-Alila, M. Coustaty, C. Faucher, and R. Wannous. Projet tourin-
flux: Apport des technologies du web sémantique pour la gestion des données
du tourisme. In 6ème édition du colloque pluridisciplinaire AsTRES: Association
Tourisme Recherche et Enseignement Supérieur, page 12, 2016.
6. F. Soualah-Alila, M. Coustaty, N. Rempulski, and A. Doucet. Datatourism: design-
ing an architecture to process tourism data. In Information and communication
technologies in tourism 2016, pages 751–763. Springer, 2016.
7. F. Soualah-Alila, C. Faucher, F. Bertrand, M. Coustaty, and A. Doucet. Applying
semantic web technologies for improving the visibility of tourism data. In Proceed-
ings of the Eighth Workshop on Exploiting Semantic Annotations in Information
Retrieval, pages 5–10. ACM, 2015.
8. CDT : Comité Départemental du Tourisme. https://fr.wikipedia.org/wiki/
Comité_départemental_du_tourisme
8