Content uploaded by Jorge Rocha
Author content
All content in this area was uploaded by Jorge Rocha on Dec 14, 2017
Content may be subject to copyright.
sustainability
Article
Identifying Tourist Places of Interest Based on Digital
Imprints: Towards a Sustainable Smart City
Luis Encalada 1ID , Inês Boavida-Portugal 2, Carlos Cardoso Ferreira 1ID and Jorge Rocha 1, *ID
1Institute of Geography and Spatial Planning, Universidade de Lisboa, Rua Branca Edmée Marques,
1600-276 Lisboa, Portugal; luisencalada@campus.ul.pt (L.E.); carlosferreira@campus.ul.pt (C.C.F.)
2Department of Spatial Planning and Environment, University of Groningen (RUG) P.O. Box 800,
9700 AV Groningen, The Netherlands; i.boavida.portugal@rug.nl
*Correspondence: jorge.rocha@campus.ul.pt; Tel.: +351-210-442-950
Received: 31 October 2017; Accepted: 9 December 2017; Published: 13 December 2017
Abstract:
As cities become increasingly complex, Information and Communication Technologies
(ICTs) bring smartness into organisations and communities, contributing to a more competitive
tourism destination, i.e., smart tourism destinations. Enhanced information access coupled with a
new kind of tourists avid for online content and predisposed to share information on social media,
allows for a better understanding of tourist behaviour regarding their spatial distribution in urban
destinations. Thus, smart tourism portrays individuals as information makers, refining the available
alternatives for tracking their location. Big data analytics is a technology with the potential to develop
Smart City services. From the analysis of the spatial distribution of tourists in the city of Lisbon based
on data collected from the ‘Panoramio’ social network, we identify the most popular places in the
city in a context of tourist visits. This new data largely contributes to understanding the consumption
of space within urban tourist destinations and therefore enables us to differentiate the overcrowded
places from the ones with potential to grow. This allows decision-makers to imagine new ways of
planning and managing towards a sustainable ‘smart’ future.
Keywords: smart city; smart tourism; big data; sustainability; social networks; geotagged photos
1. Introduction
1.1. ICTs and Big Data Analytics for Rising Smartness on City Tourism Destinations
According to the United Nations, by the year 2050 almost 66% of the world’s population will be
living in cities [
1
]. This urban pressure will raise a variety of problems and important challenges that
will likely question the economic, social, and environmental sustainability of cities [2,3].
In parallel with this inexorable and widespread population concentration in urban centres, cities
are also the object of growing tourism production and consumption, capturing ever-increasing flows of
visitors, according to several authors [
4
], as well as recent reports and insights (e.g., Global Destination
Cities Index [5]; Euromonitor International [6]).
Tourism is not only a mobility pattern in the cities and thus a geographical feature in itself, but it
is also “
. . .
assuming a growing political and cultural importance indirectly in relation to the ranking
of world cities as places to live, work and do business” [
7
]. As UNWTO claims, on a wider perspective,
tourism is becoming “
. . .
a very important element in all policies related to urban development (
. . .
)
not just a strategy to provide a competitive product to meet visitors’ expectations but a way to develop
the city itself and provide more and better infrastructures and bring conditions to residents” [8].
With more than half of the world’s population living in cities, infrastructure will face increasing
pressure [
9
]. In particular, non-digital and non-automated infrastructures of the city are already dealing
with several technical and physical problems [
10
]. Nowadays, the performance of cities depends
Sustainability 2017,9, 2317; doi:10.3390/su9122317 www.mdpi.com/journal/sustainability
Sustainability 2017,9, 2317 2 of 19
not only on their physical infrastructure, but also on accessibility and the quality of knowledge
communication and social networks [11].
Smart cities have been presented as a panacea to urbanization driven problems and a way to
achieve sustainable development [
12
]. Information and Communication Technologies (ICTs) are in
the core of this discourse [
13
], which underlines the enhancement of cities’ performance in many
fields [14].
Supporters of the smart cities paradigm stress the potential for stimulating ecological integrity
and social equity towards the greater aim of urban sustainability [
15
]. Opponents argue that slight
attention is given to the social aspects of sustainable development and to the basic social dimension
of the city [
16
]. So far, smart cities enthusiasts have succeeded in hiding answers to the big issues
(
e.g., sustainability
) rather than providing information [
17
]. Still, smart cities are a strategic vision to
stretch for sustainable futures, rather than a depiction of the current reality [18].
As cities become increasingly complex systems, ICTs intertwine all activities and services, fostering
more connected, informed, and involved citizens. ICTs make cities more accessible and pleasant for
both residents and visitors, and help cities to manage the challenges that keep emerging [19,20].
ICTs have also brought several new tools to the tourism industry. Nowadays, tourism is facing a
new set of challenges resulting from modifications in tourism consumers’ behaviour by influence of
the emerging ICTs and by the generalisation of their use as well [
21
]. From the standpoint of tourism,
ICTs can potentially contribute by generating added-value experiences for tourists, while refining
efficiency and supporting automated processes [22].
The smart city paradigm perfectly embraces ICTs as it is characterised by an environment where
technology is rooted in the city [
23
] and facilitates access to enhanced services for city visitors (both
residents and tourists), such as access to real-time information. Smart cities include different areas
of the city administration. For example, the concept of smart tourism destination arise from the
development of smart cities [
24
], i.e., the blossoming of the smart city also encourages the emergence
of smart tourism.
The World Tourism Organization (UNWTO) introduced the smart tourism concept [
25
] and
defined it as clean, green, ethical, and quality tourism, among other characteristics. Thus, smart
tourism should be capable of meeting the requirements of both short-term economic needs and
long-term sustainable development. Since tourism is a complex activity, smartness applied to tourism
should encompass both the availability of services and the efficiency of the city as a whole. Smart
tourism destinations can be perceived as places utilizing the ICTs to co-create value, pleasure, and
experiences for the tourist and profit for the destination administration [24].
From a technological point of view, smart city and smart tourism are two strongly linked
notions [
19
]. Smart tourism refers to the use of technology, attributing more importance to the
number and/or the quality of tourism-related apps available in the city, while the tourism smartness
paradigm has to account for the temporary dimension of tourism practiced by non-residents [19,26].
Yet, the integration of ICTs per se within a tourism destination is not enough for a city to become
a smart tourism destination. There is a need for enhancing human capital and other forms of skill
development among the citizenry [
3
]. Destination managers have to understand the multifaceted
genesis of smartness in order to enhance competitiveness [27].
When it comes to creating sustainable smart cities, it is imperious to understand what makes a
city tick and what makes people want to go there. Therefore, becoming smart implies reinforcing a
city’s uniqueness rather than allowing it to become impersonal and homogenized. ICTs must increase
efficiency but at the same time improve the liveability of the city, promoting engagement with the city
and respect for the environment and the several nuances of a city’s cultural identity, i.e., what makes it
different from other cities.
Bakhtin [
28
] coins three terms to analyse the clash between tourists, residents and local cultures,
namely heteroglossia, polyphony and carnivalesque. The first one focuses on understanding tourism
and its impact on local cultures. The second term stresses the ubiquitous and mediating mission of
Sustainability 2017,9, 2317 3 of 19
the tourism authorities, i.e., operators, tourists, and local residents contribute to creating the profile of
each destination, but the authorities have the responsibility of defining the overall strategy to enhance
marketability [
29
]. Finally, the third term accounts for the multiplicity of situations beyond the control
of the authorities. The way that attractions selection become hot (or cold) on social media informs the
choices that authorities have to make in their strategies.
Furthermore, the active role of the anthropic system, i.e., the human factor represented by
city users—both residents and tourists—has been gradually gaining importance as it bears direct
implications on the way cities become smart.
In fact, as stated by some authors [
26
,
27
,
30
], it is the interconnectivity between heterogeneous
human actors and the digital ecosystem—more precisely, the co-created value collaboratively shaped
by this socio-technical synergy—that contributes to introducing smartness in cities. Hence, tourism
can be a privileged field to test actual accomplishment of the potentials made available by the smart
city paradigm [25,31].
The smart city paradigm and consequently the smart tourism approach have to address both
tourists’ behaviour and consumption. These should be sustainable and restrained, preventing the
urban system from exceeding its carrying capacity threshold and degenerating into an entropic
region [31].
Some, although not all, of the smart environment features foster improved quality of life,
governance, resilience and principally an intelligent managing of city facilities and natural resources,
and thus sustainability [32].
1.2. Research Framework
Smart Tourism Destination initiatives rely on the combination of hard and soft smartness
components [
27
]. While the former stress recent developments in ICTs and the Internet of Things (IoT),
both have empowered the gathering and analysis of a huge amount of information (referred to as big
data), providing real-time insights on human behaviour [
33
]. Only in association with the anthropic
system can big data provide value for a competitive advantage of tourism destinations.
Hence, the deployment of ICTs has become a fundamental infrastructure that contributes to
competitiveness. The latest developments in Social Network platforms and the IoT are producing new
impacts on tourism destinations and, particularly, supporting information exchange regarding tourism
activities [34,35].
One of the great challenges of smart cities is the mining of relevant information from the ICTs
infrastructure of cities. Such extraction usually relies on the use of sensors that require high public
investment [
36
]. To overcome such a difficulty, some studies suggest using social media to identify the
perception of residents and visitors about a particular city [11,19,27].
Big data analytics has emerged as a technology with the potential to augment smart city
services [
32
,
37
]. Big data is characterised by a huge volume and variety of data types; created at
continually increasing rates [
38
]. Big data analytics allows for the extraction of meaningful information
from large amounts of data produced by ICTs [37–39].
Big data provides insights from diverse sources of data. The features derived from social media
are typically unstructured when compared with big data collected in other contexts [40].
The better access to information facilitated by ICTs and a new profile of tourists seeking and
sharing online content have contributed to an improved knowledge of the characteristics and behaviour
of tourists [
41
,
42
]. In this context, individuals themselves have become generators of information,
contributing to boost the available alternatives for tracking their location [43].
User-generated data retrieved from social media include, among others, geotagged photos.
Contrasting traditional survey-based methods, this data can offer cost-effective information on
activities and preferences of tourists [
44
]. Social media has become gradually more important in
many aspects of tourism, particularly in information search, decision-making behaviours, tourism
promotion and in focusing on best practices for interacting with consumers [45].
Sustainability 2017,9, 2317 4 of 19
All this innovation generated by ICTs in terms of information and information sources has
emerged as an additional and complementary support tool for the more formal/traditional sources of
data collection in tourism, particularly in local contexts (e.g., urban destinations) where there are still
various constraints on the availability and appropriateness of information [7].
Moreover, it could also suppress restrictions that came from problems like sample size,
nonresponse bias, self-reported errors and time and location constrains [
46
]. Nonetheless, there
are still some concerns regarding the use of social media content in this context [
47
]. These are mainly
related to data quality [
33
], to the posts potential location inaccuracy, to the representativeness of the
population using social media [48] and to the possible biased behaviour on social media [49].
When tourists use their credit cards or mobile phones, and share or post content in social networks,
they leave multiple digital footprints that result in a multidimensional data set [
19
]. Considering
limitations about collecting data to study urban tourism [
7
], this information can be used as a proxy
to measure the attractiveness of places as well as the spatial distribution of tourists in the city [
41
,
50
].
Different types of tourists have different behaviours in using social media and different social media
users have different impacts on potential tourists [45].
For instance, from the visitor’s perspective, smart user oriented location-based services can lead
tourists to the most attractive places—i.e., hotspots [
51
,
52
]. Social media can reflect and influence
tourists
'
perceptions and approaches concerning green/eco actions. Social media users are expected
to be aware of the impact of travel behaviours and to look for the most sustainable alternatives [
45
].
From the city’s perspective, management information systems—i.e., business intelligence—allows
the distributed generation of specific knowledge, relevant to those that manage destinations and to
private/public stakeholders [53,54].
Some empirical studies have led to identifying the spatial patterns of city tourists by analysing
data from social media [
55
–
57
]. However, few attempts have explored the factors that can explain
this observed spatial pattern. Thus, the aim of this study is to (1) analyse the spatial distribution of
tourists in Lisbon based on data collected from the ‘Panoramio’ social network, and (2) to explore
some relations between the observed pattern and a set of variables related to the city tourism offer.
This data embedded in the new ICT context supports the exploration of patterns, while contributing to
the understanding of urban tourism spatial production and consumption.
The geotagged photos published by tourists on ‘Panoramio’—based on visits from 2008 to
2014—have allowed us to reach a quantitative and geographic perspective on the city’s tourist
consumption: Which are the most visited city areas? What is their location in the urban context?
From the analysis of a set of online tourist guides related to the city of Lisbon, it is also possible
to identify the most referred locations for tourists. This shows us how Lisbon is presented as a
tourist destination and the differentiated attractiveness of the places to visit, based on the indications
and preferences from tourist online guides that hence constitute an approach from the city
'
s tourist
production perspective.
2. Materials and Methods
2.1. Study Area
Centrally located in the metropolitan area of Lisbon (LMA), the city of Lisbon (Figure 1) has an
area of about 85 km
2
and 547,773 inhabitants [
58
]. Lisbon is regarded as the urban centrality of the
LMA according to the Strategic Plan for Tourism in the Lisbon Region (PETRL) 2015–2019 [
59
]; the city
is one of the tourist regional centres, showcasing itself as a strong international brand that acts as an
anchor of the remaining centralities (Cascais, Sintra, Arrábida, and Arco do Tejo [Tagus Axe]).
Regarding tourism demand statistics, in 2016, the LMA concentrated 24.9% of total overnight
stays in the country. The region continues to be among the first choices for major external markets,
being the main destination for Brazilians, North Americans, and Italians. Concerning the internal
market, it is the second destination of choice for nationals (with 19.8% of total overnight stays) [60].
Sustainability 2017,9, 2317 5 of 19
In 2014, the average stay of foreign guests in the city of Lisbon was of 2.5 nights, below the average
of the LMA (2.6 nights) and mainland Portugal (3.1 nights) [61].
Sustainability 2017, 9, 2317 5 of 19
Figure 1. Lisbon Region Centralities [59].
Some areas of the city contribute most to its overall touristic appeal. These areas—also called
micro-centralities (MC), following the proposed zoning in the PETRL 2015–2019 (Figure 2 and Table
1)—are situated in three different locations, from the southwest, in Belém, crossing the historic
nucleus, in the city centre, to the northeast part of the city, in Parque das Nações.
Figure 2. Tourism Micro-centralities in Lisbon [59].
According to the survey of tourist’s motivations [62] and the survey of visitor’s activities [63]—
both conducted in 2014 by the Tourism Observatory of Lisbon at the international and national
Figure 1. Lisbon Region Centralities [59].
Some areas of the city contribute most to its overall touristic appeal. These areas—also called
micro-centralities (MC), following the proposed zoning in the PETRL 2015–2019 (Figure 2and
Table 1)—are situated in three different locations, from the southwest, in Belém, crossing the historic
nucleus, in the city centre, to the northeast part of the city, in Parque das Nações.
Sustainability 2017, 9, 2317 5 of 19
Figure 1. Lisbon Region Centralities [59].
Some areas of the city contribute most to its overall touristic appeal. These areas—also called
micro-centralities (MC), following the proposed zoning in the PETRL 2015–2019 (Figure 2 and Table
1)—are situated in three different locations, from the southwest, in Belém, crossing the historic
nucleus, in the city centre, to the northeast part of the city, in Parque das Nações.
Figure 2. Tourism Micro-centralities in Lisbon [59].
According to the survey of tourist’s motivations [62] and the survey of visitor’s activities [63]—
both conducted in 2014 by the Tourism Observatory of Lisbon at the international and national
Figure 2. Tourism Micro-centralities in Lisbon [59].
Sustainability 2017,9, 2317 6 of 19
According to the survey of tourist’s motivations [
62
] and the survey of visitor’s
activities [
63
]—both conducted in 2014 by the Tourism Observatory of Lisbon at the international and
national levels—‘City & short break’ is considered the largest motivation for visiting Lisbon. For about
two-thirds of those surveyed, the city of Lisbon was their only travel destination (64.1% of these were
first-comers). Visiting monuments, attractions or museums was their preferred activity, particularly in
the case of ‘First-comers’ (81.3%) in comparison with repeaters (66.3%).
There is a predominant interest for sites within the city, and those located in the MC Belém and
Baixa Chiado had the highest number of visits (Table 1). From these surveys, the importance of the
historical-monumental nucleus is also clear, from which Belém and Saint Jorge Castle stand out as the
most visited attractions.
Table 1. Places of interest and visited attractions—Survey of visitors’ activities [63].
MC Places of Interest % Places Visited % MC
BE Belém 79.8 Belém Tower 87.1 BE
BC Baixa 74.9 Jerónimos Monastery 71.9 BE
PN Parque das Nações 72.6 Saint Jorge Castle 63.5 ACM
BC Terreiro do Paço 72.1 Monument of the Discoveries 63.1 BE
PL Avenida da Liberdade 70.3 Lisbon Oceanarium 40.6 PN
BCS Bairro Alto 67.2 Lisbon Cathedral (Sé) 37.1 ACM
ACM Alfama 64.8 Ajuda Palace 28.8 BE
BC Chiado 57.0 Pavilion of Knowledge 23.8 PN
Belém Cultural Centre 21.6 BE
Gulbenkian Museum 14.7 PL
Lisbon Casino 11.8 PN
Ancient Art Museum 10.5 BCS
Coaches Museum 10.4 BE
Chiado Museum 8.6 BC
Design Museum (MUDE) 2.8 BC
Lisbon Story Centre 2.2 BC
2.2. Data Retrieving Process and Data Aggregation
In order to retrieve the photos from ‘Panoramio’, the study area was segmented into smaller
subareas. For each new segment, we extracted the information of the posted photos within its extent.
First, the study area was segmented into a lattice of 1000 m
×
1000 m squares. This resulted in
49,048 records
. This process was subsequently repeated with increasingly smaller tiles (reaching
22 iterations
) until the number of retrieved photos stabilised, i.e., the last tests resulted in a similar
number of photos. The final database counted on more than 70,000 records.
The method to differentiate the photos taken by visitors from the ones taken by locals was based
on the work carried out by Girardin et al. [
55
], Kádár [
56
], and García-Palomares et al. [
57
]. The authors
calculated the difference (in days) between the timestamp of the first and the last uploaded photo by
each user. If the resulting number exceeded the average visit time associated with the destination, the
photos were considered as belonging to locals; conversely, if the number was beyond the time-period,
the photos were catalogued as belonging to visitors.
The resulting dataset consists of 17,604 uploaded photos taken by users that were considered
visitors. The location of those photographs is depicted in Figure 3. At first glance, some areas are
clearly more prominent than others, i.e., they present a higher concentration of uploaded photos, thus
highlighting the most visited/photographed places in the city.
Sustainability 2017,9, 2317 7 of 19
Sustainability 2017, 9, 2317 7 of 19
Figure 3. Photos taken in Lisbon between 2008 and 2014, and uploaded on ‘Panoramio’.
For the statistical analysis, the dataset was aggregated from a point layer to a regular grid of
hexagons with ≈2500 m
2
each. The value assigned to each of the hexagonal units corresponds to the
sum of the individual observations within each cell (Figure 4).
For the multiple regression analysis, the data was aggregated considering the basic units for
census purposes, called ‘subsections’.
Figure 4. Aggregated tourist photos, taken in Lisbon between 2008 and 2014 and uploaded on
‘Panoramio’. Boxes A (Belém) and B (Historic centre) show zoom frames of the two most visited areas.
Figure 3. Photos taken in Lisbon between 2008 and 2014, and uploaded on ‘Panoramio’.
For the statistical analysis, the dataset was aggregated from a point layer to a regular grid of
hexagons with
≈
2500 m
2
each. The value assigned to each of the hexagonal units corresponds to the
sum of the individual observations within each cell (Figure 4).
For the multiple regression analysis, the data was aggregated considering the basic units for
census purposes, called ‘subsections’.
Sustainability 2017, 9, 2317 7 of 19
Figure 3. Photos taken in Lisbon between 2008 and 2014, and uploaded on ‘Panoramio’.
For the statistical analysis, the dataset was aggregated from a point layer to a regular grid of
hexagons with ≈2500 m
2
each. The value assigned to each of the hexagonal units corresponds to the
sum of the individual observations within each cell (Figure 4).
For the multiple regression analysis, the data was aggregated considering the basic units for
census purposes, called ‘subsections’.
Figure 4. Aggregated tourist photos, taken in Lisbon between 2008 and 2014 and uploaded on
‘Panoramio’. Boxes A (Belém) and B (Historic centre) show zoom frames of the two most visited areas.
Figure 4.
Aggregated tourist photos, taken in Lisbon between 2008 and 2014 and uploaded on
‘Panoramio’. Boxes A (Belém) and B (Historic centre) show zoom frames of the two most visited areas.
Sustainability 2017,9, 2317 8 of 19
2.3. Cluster and Outlier Analysis
The Cluster and Outlier analysis is based on the Local Moran Index [
64
], which enables the
identification of local patterns of spatial association. This indicator distinguishes the relationship
between the attribute value of an entity and the values corresponding to the entities in its
neighbourhood, and is also able to identify the extent of the spatial concentration of similar values for
each observation.
In order to perform the analysis, it is necessary to define the neighbourhood area (area surrounding
entities to compare) for each target feature and the nature of the spatial relationship between entities.
The neighbourhood was represented on a spatial weight matrix. Following Tobler’s first Law of
Geography, which states that everything is related to everything but near things are probably more
related that distant ones, the inverse distance was chosen to conceptualise the relationship between
spatial observations. Thus, the influence of the neighbouring features will decrease as the distance
between them increases. Assuming that not all uploads within the study area are spatially related, a
threshold for the neighbourhood radius of influence was determined (150 m).
2.4. Multiple Linear Regression
In order to explore the relationships between the observed pattern of tourists’ photos and a set of
24 selected variables related to the city
'
s tourist offer, e.g., attractions, services and facilities, a multiple
linear regression analysis was carried out. One of the first things to consider in the implementation
of a linear regression model is the selection of explanatory factors (variables). A single additional
independent variable can improve the prediction of the dependent variable and therefore they must
be chosen carefully. Here, 24 variables were selected to be used as independent variables (Table 2), i.e.,
as explanatory factors of the geographical agglomeration of photos.
Table 2. Pre-selected explanatory variables.
Code Variable Code Variable
V1 Civil architecture V13 Hospitality (Accommodation)
V2 Industrial architecture V14 Churches
V3 Military architecture V15 Markets
V4 Noble architecture V16 Viewpoints
V5 Award-winning architecture V17 Monuments of public interest
V6 Religious architecture V18 National monuments
V7 Cemeteries V19 Museums
V8 Blocks of public interest V20 Playgrounds
V9 Leisure docks and marinas V21 Picnic parks
V10 Funiculars and lifts V22 Leisure parks
V11 Fairs V23 Urban parks
V12 Geo-monuments V24 Theatres
Spatially, all these variables were represented in a point layer. Subsequently, all variables were
transformed into reason scales by applying the Inverse Euclidean Distance to the parsed element. As
the unit of measurement is always the same (metres), there was no need to standardise values. In order
to illustrate the transformation of the variables, the map with the Inverse Euclidean Distance to the
features of the variable ‘Viewpoints’ (V16) is presented in Figure 5.
The predictive performance of a linear regression model is not only related to the correlation of
the dependent variable with a certain (independent) variable, but also to the correlation of the former
with all the other independent variables. Thus, the possible dependencies among all the variables
should be assessed. In cases where these dependencies between variables are strong, one can state that
there is a multicollinearity process.
Sustainability 2017,9, 2317 9 of 19
Sustainability 2017, 9, 2317 9 of 19
Figure 5. Inverse Euclidean Distance to viewpoint features.
As collinearity means that some independent variables are correlated, it can be detected, among
different ways, through the array of correlations that exist between them. The Pearson correlation
coefficients between independent variables were calculated. Only variables with r > −0.85 or r < 0.85
correlations were included in the regression model without the risk of high multicollinearity [65].
Looking at Figure 6, one can see that strong positive relations are predominant, whereas negative
relations are only moderate.
Figure 6. Pearson correlation coefficient between the 24 pre-selected explanatory variables.
Figure 5. Inverse Euclidean Distance to viewpoint features.
As collinearity means that some independent variables are correlated, it can be detected, among
different ways, through the array of correlations that exist between them. The Pearson correlation
coefficients between independent variables were calculated. Only variables with r>
−
0.85 or r< 0.85
correlations were included in the regression model without the risk of high multicollinearity [
65
].
Looking at Figure 6, one can see that strong positive relations are predominant, whereas negative
relations are only moderate.
Sustainability 2017, 9, 2317 9 of 19
Figure 5. Inverse Euclidean Distance to viewpoint features.
As collinearity means that some independent variables are correlated, it can be detected, among
different ways, through the array of correlations that exist between them. The Pearson correlation
coefficients between independent variables were calculated. Only variables with r > −0.85 or r < 0.85
correlations were included in the regression model without the risk of high multicollinearity [65].
Looking at Figure 6, one can see that strong positive relations are predominant, whereas negative
relations are only moderate.
Figure 6. Pearson correlation coefficient between the 24 pre-selected explanatory variables.
Figure 6. Pearson correlation coefficient between the 24 pre-selected explanatory variables.
Another technique used was the tolerance value or its inverse, known as Variance Inflation Factor
(VIF), whose high values indicate the existence of multicollinearity. The VIF is a measure of the degree
Sustainability 2017,9, 2317 10 of 19
to which each independent (temporarily transformed in dependent) variable is explained by all the
other independent variables [
66
]. The greater the VIF, the more severe will be the multicollinearity.
Thus, the VIF measures the redundancy between all explanatory variables. Explanatory variables with
VIF values above 7.5 were removed (one by one) from the regression model [
66
,
67
]. Table 3shows VIF
values for each of the 24 pre-selected variables. Of these, 12 were excluded for presenting VIF values
above the defined threshold.
Table 3. Variance Inflation Factor (VIF) and frequency values.
Code Variable VIF Frequency
V23 Urban parks 2.99 0
V7 Cemeteries 3.50 0
V17 Monuments of public Interest 3.96 1
V20 Playgrounds 4.14 3
V24 Theatres 4.27 1
V5 Award-winning architecture 4.50 1
V1 Civil architecture 5.39 0
V12 Geo-monuments 5.70 21
V2 Industrial architecture 5.87 0
V13 Hospitality (Accommodation) 6.16 33
V19 Museums 6.25 0
V9 Leisure docks and marinas 6.70 6
V15 Markets 10.31 8
V4 Noble architecture 16.83 1
V14 Churches 16.87 32
V3 Military architecture 20.85 0
V11 Fairs 22.37 0
V6 Religious architecture 45.08 0
V16 Viewpoints (Belvederes) 58.24 2
V18 National monuments 58.77 0
V8 Blocks of public interest 104.93 0
V10 Funiculars and lifts 114.61 29
V21 Picnic parks 359.06 5
V22 Leisure parks 442.87 5
Complementarily, an exploratory analysis was performed in which the dependent variable was
modelled against sets of 1 to 5 independent variables. Overall, 35,526 combinations were tested, 6866
of which presented an R
2
greater than 0.90. Within this group, the 38 best performing models always
used 5 variables, with predictive performances between 0.95 and 0.96. Table 3also identifies the
variables that were most frequently selected in the testing models.
The cross-tabulation between VIF values and frequencies (Table 3), i.e., variables with a higher
frequency but with low VIF values, guided the final selection of 7 independent variables to customise
the model (Table 4).
Table 4. Multiple Linear Regression final variables selection.
Code Variable Weight
V13 Hospitality 0.10
V12 Geo-monuments 0.12
V9 Leisure docks and marinas 0.21
V20 Playgrounds 0.03
V5 Award-winning architecture 0.15
V17 Monuments of public interest 0.29
V24 Theatres 0.10
Sustainability 2017,9, 2317 11 of 19
R
2
and adjusted R
2
measures were verified as performance indicators of the model. As a final
result, the regression achieved values of R
2
= 0.6828 and adjusted R
2
= 0.6827, which means the model
is able to represent about 68% of the spatial variation of the total amount of tourists’ photos.
3. Results
3.1. Analysis by Micro-Centralities
Taking into account the micro-centralities defined according to the zoning proposed in the
2015–2019 PETRL, it is possible to identify tourist preferences regarding some places of the city that are
perceived as areas of interest (Figure 7). In fact, 75% of the photos were taken within the boundaries
of those areas defined as micro-centralities. This number is divided between the historic centre, the
main focus of tourists’ photos (more than 40% of overall photos), Belém (about 20%), and Parque das
Nações (with 12%).
Sustainability 2017, 9, 2317 11 of 19
Table 4. Multiple Linear Regression final variables selection.
Code Variable Weight
V13 Hospitality 0.10
V12 Geo-monuments 0.12
V9 Leisure docks and marinas 0.21
V20 Playgrounds 0.03
V5 Award-winning architecture 0.15
V17 Monuments of public interest 0.29
V24 Theatres 0.10
3. Results
3.1. Analysis by Micro-Centralities
Taking into account the micro-centralities defined according to the zoning proposed in the 2015–
2019 PETRL, it is possible to identify tourist preferences regarding some places of the city that are
perceived as areas of interest (Figure 7). In fact, 75% of the photos were taken within the boundaries
of those areas defined as micro-centralities. This number is divided between the historic centre, the
main focus of tourists’ photos (more than 40% of overall photos), Belém (about 20%), and Parque das
Nações (with 12%).
Despite the predominance of the historic centre, the distribution of the photographs within its
limits is geographically asymmetric and the predominance of two subareas over the others can be
clearly noted. In this context, MC I (Baixa-Chiado) and II (Alfama/Castelo/Mouraria), with
respectively 35% and 32% of the total number of photos taken in the historic centre, differ from the
others, MC III (Avenida da Liberdade) with 16% and MC IV (Bairro Alto/Santos) with 15%.
Figure 7. Number of photos per micro-centralities.
3.2. Clusters and Outlier Analysis
As expected, clusters are located within the tourist areas (MC) of the city, mostly neighbouring
the most relevant attractions. As shown in Section 3.1, when analysing the absolute number of photos,
the well-known city tourist areas understandably stand out. However, this approach did not include
some key aspects in regards to spatial analysis, as the configuration of the local neighbourhood or
the statistical significance of the data.
Figure 7. Number of photos per micro-centralities.
Despite the predominance of the historic centre, the distribution of the photographs within
its limits is geographically asymmetric and the predominance of two subareas over the others can
be clearly noted. In this context, MC I (Baixa-Chiado) and II (Alfama/Castelo/Mouraria), with
respectively 35% and 32% of the total number of photos taken in the historic centre, differ from the
others, MC III (Avenida da Liberdade) with 16% and MC IV (Bairro Alto/Santos) with 15%.
3.2. Clusters and Outlier Analysis
As expected, clusters are located within the tourist areas (MC) of the city, mostly neighbouring
the most relevant attractions. As shown in Section 3.1, when analysing the absolute number of photos,
the well-known city tourist areas understandably stand out. However, this approach did not include
some key aspects in regards to spatial analysis, as the configuration of the local neighbourhood or the
statistical significance of the data.
As shown in Figure 8, the most striking fact is that almost all of the significant spatial clusters
belong to the High-High (HH) category, i.e., they correspond to places where high values are
surrounded by similar values. Therefore, there is a large presence of visitors, both in these locations
and in their contiguous areas.
Sustainability 2017,9, 2317 12 of 19
Sustainability 2017, 9, 2317 12 of 19
As shown in Figure 8, the most striking fact is that almost all of the significant spatial clusters
belong to the High-High (HH) category, i.e., they correspond to places where high values are
surrounded by similar values. Therefore, there is a large presence of visitors, both in these locations
and in their contiguous areas.
Figure 8. Clusters and outliers of tourists’ photos in Lisbon. Boxes A (Belém) and B (Historic centre)
show zoom frames of the two most visited areas.
On the other hand, only two clusters represent a different category—Low-High (LH)—and refer
to those locations which specifically do not have that many visitors but that are characterized by
significant numbers of visitors in their neighbourhoods.. Therefore, these particular spots are less
visited than the areas surrounding them.
In fact, it turns out that tourist attractions—e.g., squares, viewpoints, monumental
architecture—function as the main focus of spatial clusters. It is around them that clusters take form,
and it is from them that these expand to nearby areas. In many cases, the clusters’ expansion follows
the physical shape of the attractions (e.g., squares), extending beyond their perimeter.
The Local Moran Index was calculated (Figure 9) to understand the difference between spatial
clusters with regards to their magnitude.
The higher values of the index are associated with locations around or in the neighbourhood of
the attractions (e.g., Belém Tower and the Monument of the Discoveries to the south, and Praça do
Marquês de Pombal in the historic centre). In the case of the Jerónimos Monastery and the Lisbon
Oceanarium, the more differentiated values correspond to places located at the front, in the area near
the entrances.
Regarding clusters on squares (e.g., ‘Restauradores’, ‘Dom Pedro IV’, ‘Comércio’ and ‘Luís de
Camões’ squares), the most predominant values of the index tend to be closer to the centre. For
instance, in the ‘Restauradores’ and ‘Dom Pedro IV’ squares, the central area clearly stands out.
Figure 8.
Clusters and outliers of tourists’ photos in Lisbon. Boxes A (Belém) and B (Historic centre)
show zoom frames of the two most visited areas.
On the other hand, only two clusters represent a different category—Low-High (LH)—and refer
to those locations which specifically do not have that many visitors but that are characterized by
significant numbers of visitors in their neighbourhoods.. Therefore, these particular spots are less
visited than the areas surrounding them.
In fact, it turns out that tourist attractions—e.g., squares, viewpoints, monumental
architecture—function as the main focus of spatial clusters. It is around them that clusters take
form, and it is from them that these expand to nearby areas. In many cases, the clusters’ expansion
follows the physical shape of the attractions (e.g., squares), extending beyond their perimeter.
The Local Moran Index was calculated (Figure 9) to understand the difference between spatial
clusters with regards to their magnitude.
The higher values of the index are associated with locations around or in the neighbourhood of
the attractions (e.g., Belém Tower and the Monument of the Discoveries to the south, and Praça do
Marquês de Pombal in the historic centre). In the case of the Jerónimos Monastery and the Lisbon
Oceanarium, the more differentiated values correspond to places located at the front, in the area near
the entrances.
Regarding clusters on squares (e.g., ‘Restauradores’, ‘Dom Pedro IV’, ‘Comércio’ and ‘Luís de
Camões’ squares), the most predominant values of the index tend to be closer to the centre. For
instance, in the ‘Restauradores’ and ‘Dom Pedro IV’ squares, the central area clearly stands out.
Sustainability 2017,9, 2317 13 of 19
Sustainability 2017, 9, 2317 13 of 19
Figure 9. Local Moran Index. Boxes A (Belém) and B (Historic centre) show zoom frames of the two
most visited areas.
3.3. Multiple Linear Regression Analysis
From the Multiple Linear Regression (MLR) analysis, one can identify the most relevant factors,
and their importance (weights), thus explaining the spatial distribution of photographs. The resulting
explanatory factors are the proximity to: Public monuments (0.29); Leisure docks and marinas (0.21);
Award-winning architecture (0.15); Geo-monuments (0.12); Hospitality (Accommodation) (0.10); and
Theatres (0.10). However, beyond the analysis of these factors and the prediction model itself, the
analysis of the residuals is also interesting for tourism planning purposes.
The residuals analysis is precisely a technique used to check the suitability of a regression model.
The basic idea of the residuals analysis is that if the model fits, the residuals should reflect the
properties of the model error, i.e., positive residuals reveal results above the predicted values and
negative residuals indicate the opposite.
In practice, what ‘residuals’ means is that, given the observed conditions—i.e., the 7 explanatory
factors—there are areas where the number of pictures is higher than expected (positive residuals)
and others with fewer pictures than would be expected (negative residuals). Therefore, those deficit
areas may require some assistance in terms of planning in order to reach their full potential.
In the south of Lisbon, Belém stands out with values above its potential (Figure 10).
Unexpectedly, the downtown area (Baixa-Chiado), albeit having appeared in all the previous
analyses as a major hot spot, is still below its potential. The Parque das Nações area follows the same
trend but without a prominent disparity. However, they are clearly two micro-centralities identified
by the Tourism Bureau that may be the subject of planning actions targeted at attracting more visitors.
Figure 9.
Local Moran Index. Boxes A (Belém) and B (Historic centre) show zoom frames of the two
most visited areas.
3.3. Multiple Linear Regression Analysis
From the Multiple Linear Regression (MLR) analysis, one can identify the most relevant factors,
and their importance (weights), thus explaining the spatial distribution of photographs. The resulting
explanatory factors are the proximity to: Public monuments (0.29); Leisure docks and marinas (0.21);
Award-winning architecture (0.15); Geo-monuments (0.12); Hospitality (Accommodation) (0.10); and
Theatres (0.10). However, beyond the analysis of these factors and the prediction model itself, the
analysis of the residuals is also interesting for tourism planning purposes.
The residuals analysis is precisely a technique used to check the suitability of a regression model.
The basic idea of the residuals analysis is that if the model fits, the residuals should reflect the
properties of the model error, i.e., positive residuals reveal results above the predicted values and
negative residuals indicate the opposite.
In practice, what ‘residuals’ means is that, given the observed conditions—i.e., the 7 explanatory
factors—there are areas where the number of pictures is higher than expected (positive residuals) and
others with fewer pictures than would be expected (negative residuals). Therefore, those deficit areas
may require some assistance in terms of planning in order to reach their full potential.
In the south of Lisbon, Belém stands out with values above its potential (Figure 10). Unexpectedly,
the downtown area (Baixa-Chiado), albeit having appeared in all the previous analyses as a major hot
spot, is still below its potential. The Parque das Nações area follows the same trend but without a
prominent disparity. However, they are clearly two micro-centralities identified by the Tourism Bureau
that may be the subject of planning actions targeted at attracting more visitors.
Sustainability 2017,9, 2317 14 of 19
Sustainability 2017, 9, 2317 14 of 19
Figure 10. Residuals of the Multiple Linear Regression model.
4. Discussion
Online photographs published by visitors make it possible to identify and analyse visual tourist
places of interest within the city. Most studies based on information gathered from social networking
sites have been limited to creating products, e.g., maps and animations. The results show that the
photographs taken by tourists reveal higher spatial concentration than those obtained by residents.
The comparison between the spatial clusters of photos and the main tourist attractions (within
the micro-centralities) indicates that the results are consistent, since the spatial clusters match the
places of the city that are known to attract more tourists.
The data set (17,604 observations used in this analysis) retrieved from the ‘Panoramio’ social
network provides a robust sample that reproduces the city tourist areas, frequently experienced by
non-locals. Although there is a growing trend among tourists to share their travel experiences
through social networks, this data only partially represents the actual tourism demand.
The spatial distribution of tourists is not homogeneous. In an attempt to clarify the factors that
stimulate the variation of the spatial process, it is confirmed that monuments of public interest
(together with six other factors) are the elements that best explain the geographical agglomeration of
photographs. Although the model reached 68% of the phenomenon explanation, the findings shown
by regression residuals should be considered, as well as their potential contribution for tourism
management. According to the regression model, there are areas within micro-centralities whose
tourism potential is underestimated. These areas may need future intervention in terms of public
facilities, activities, or services. Since this is an empirical approach to explore the potential relations
between a set of possible factors associated to the tourist offer, further explanatory research has to be
conducted to test the model resulting factors and their level of importance (weight).
The definition of tourist areas can largely benefit from the Tourism 2.0 potential, namely by
analysing the concentration of the tourists’ digital imprints through location-based information, such
Figure 10. Residuals of the Multiple Linear Regression model.
4. Discussion
Online photographs published by visitors make it possible to identify and analyse visual tourist
places of interest within the city. Most studies based on information gathered from social networking
sites have been limited to creating products, e.g., maps and animations. The results show that the
photographs taken by tourists reveal higher spatial concentration than those obtained by residents.
The comparison between the spatial clusters of photos and the main tourist attractions (within the
micro-centralities) indicates that the results are consistent, since the spatial clusters match the places of
the city that are known to attract more tourists.
The data set (17,604 observations used in this analysis) retrieved from the ‘Panoramio’ social
network provides a robust sample that reproduces the city tourist areas, frequently experienced by
non-locals. Although there is a growing trend among tourists to share their travel experiences through
social networks, this data only partially represents the actual tourism demand.
The spatial distribution of tourists is not homogeneous. In an attempt to clarify the factors
that stimulate the variation of the spatial process, it is confirmed that monuments of public interest
(together with six other factors) are the elements that best explain the geographical agglomeration
of photographs. Although the model reached 68% of the phenomenon explanation, the findings
shown by regression residuals should be considered, as well as their potential contribution for tourism
management. According to the regression model, there are areas within micro-centralities whose
tourism potential is underestimated. These areas may need future intervention in terms of public
facilities, activities, or services. Since this is an empirical approach to explore the potential relations
between a set of possible factors associated to the tourist offer, further explanatory research has to be
conducted to test the model resulting factors and their level of importance (weight).
The definition of tourist areas can largely benefit from the Tourism 2.0 potential, namely by
analysing the concentration of the tourists’ digital imprints through location-based information, such
Sustainability 2017,9, 2317 15 of 19
as photos shared in social networks. It is possible to highlight which areas are more attractive to
visitors and, more importantly, their relevant features. Yet, some areas may not be regarded as tourist
attractions by local authorities despite their high potential for tourist activities. Knowing which areas
are under tourist pressure effectively contributes to tourism and city management and competitiveness,
providing decision-makers with improved tools to design better, smarter, and sustainable strategies;
also contributes to optimise tourists’ experience, which should be, ultimately, the goal of a smart
tourism destination.
5. Conclusions
The widespread use of ICTs—such as cloud computing with high processing performance, the
IoT, and data mining—has created different dimensions for tourism. Smartness is a key factor in
achieving sustainable development [
68
] and contributes to advances in tourism [
19
,
27
], but also
in society as a whole. Indeed, the added value of smart tourism is innovation, technology, and
sustainability [26,27,30].
The smart city concept therefore encompasses a wide application of sustainability principles to
urban competitiveness [
18
], referring both to the use of ICTs and to the quality of the soft smartness
components [20].
Since the smart city approach is grounded on the assumption that technology is part of the
system, the use of technological solutions in decision-making processes based on evidence effectively
contributes to ranking necessary actions and predicting future scenarios, which is crucial for a
responsible, smart, and sustainable management of tourism and its impacts. Thanks to these new
solutions involving smart tools, cities can improve access to tourist spots, and explore the prospects and
challenges presented by crowdsourced and open data platforms. Smart tourism is a key contributor to
a new “sensor society” [69] that is regarded as always based on extensive (big) data capture.
ICT platforms, sensor networks, and wireless communication systems form the basis for an
integration and data exchange approach [
30
] and contribute to the overall strategic planning process
by facilitating the transformation of tacit knowledge into explicit organised knowledge [
70
]. However,
the integration of technology within a tourism destination is not the sole condition for becoming
smart, and thus the goal is to integrate the whole range of smartness components. For that, ICTs
should be complemented with information provided by experts, including government officials and
researchers, in order develop an evaluation system for quality management in tourist attractions [
71
].
Nevertheless, smart tourism draws attention to some issues regarding information governance [
72
]
and appropriately extracting the value of information. Another issue is technology-dependence. This
has strong consequences in terms of creating a wide digital split between those with or without smart
devices, and between destinations that can or cannot afford smart tourism infrastructures [
73
]. In this
sense, smart tourism infrastructures can lead to new information imbalances [74].
Furthermore, the growing need for information can easily persuade tourists to lose privacy [
75
].
Smart tourists leave massive digital footprints and the chances for mining their digital traces from
travel (whether on holiday or business) are multiple. Therefore, smart tourism environments have
to consider the value-in-use [
76
], i.e., the value created by using data and/or technology instead of
owning it.
Acknowledgments:
Luis Encalada was financed through a grant of the Institute of Geography and Spatial
Planning and Universidade de Lisboa, IGOT-UL (BD2016).
Author Contributions:
Luis Encalada, Carlos Cardoso Ferreira and Jorge Rocha analysed the data; Luis Encalada,
Inês Boavida-Portugal and Jorge Rocha contributed to the analysis tools; Luis Encalada, Inês Boavida-Portugal,
Carlos Cardoso Ferreira and Jorge Rocha wrote the paper.
Conflicts of Interest: The authors declare no conflict of interest.
Sustainability 2017,9, 2317 16 of 19
References
1.
United Nations. World Urbanization Prospects: The 2014 Revision; ST/ESA/SER.A/366; United Nations:
New York, NY, USA, 2015.
2.
Organisation for Economic Co-operation and Development (OECD). OECD Environmental Outlook to 2050.
The Consequences of Inaction; OECD: Paris, France, 2012; ISBN 9789264122161.
3.
Neirotti, P.; De Marco, A.; Cagliano, A.C.; Mangano, G.; Scorrano, F. Current trends in Smart City initiatives:
Some stylised facts. Cities 2014,38, 25–36. [CrossRef]
4.
Jurdana, D.S.; Sušilovi´c, Z. Planning city tourism development: Principles and issues. Tour. Hosp. Manag.
2006,12, 135–144.
5.
Hedrick-Wong, Y.; Choong, D. MasterCard 2014 Global Destination Cities Index; MasterCard: Purchase, NY,
USA, 2014.
6.
Euromonitor International. Top 100 City Destinations Ranking; Euromonitor International: London, UK, 2014.
7.
Ashworth, G.; Page, S.J. Urban tourism research: Recent progress and current paradoxes. Tour. Manag.
2011
,
32, 1–15. [CrossRef]
8.
World Tourism Organization. Global Report on City Tourism—Cities 2012 Project; World Tourism Organization:
Madrid, Spain, 2012.
9.
Falconer, G.; Mitchell, S. Smart City Framework: A Systematic Process for Enabling Smart+ Connected Communities;
Cisco Internet Business Solutions Group (IBSG): San Jose, CA, USA, 2012.
10.
Colldahl, C.; Frey, S.; Kelemen, J.E. Smart Cities: Strategic Sustainable Development for an Urban World; Blekinge
Institute of Technology: Karlskrona, Sweden, 2013.
11. Caragliu, A.; Del Bo, C.; Nijkamp, P. Smart Cities in Europe. J. Urban Technol. 2011,18, 65–82. [CrossRef]
12. Datta, A. A 100 smart cities, a 100 utopias. Dialogues Hum. Geogr. 2015,5, 49–53. [CrossRef]
13.
Kummitha, R.K.R.; Crutzen, N. How do we understand smart cities? An evolutionary perspective. Cities
2017,67, 43–52. [CrossRef]
14.
Kourtit, K.; Nijkamp, P. Smart cities in the innovation age. Innov. Eur. J. Soc. Sci. Res.
2012
,25, 93–95.
[CrossRef]
15.
Gibbs, D.; Krueger, R.; MacLeod, G. Grappling with Smart City Politics in an Era of Market Triumphalism.
Urban Stud. 2013,50, 2151–2157. [CrossRef]
16.
Cugurullo, F. How to Build a Sandcastle: An Analysis of the Genesis and Development of Masdar City.
J. Urban Technol. 2013,20, 23–37. [CrossRef]
17.
Calzada, I.; Cobo, C. Unplugging: Deconstructing the Smart City. J. Urban Technol.
2015
,22, 23–43. [CrossRef]
18. Angelidou, M. Smart cities: A conjuncture of four forces. Cities 2015,47, 95–106. [CrossRef]
19.
Buhalis, D.; Amaranggana, A. Smart Tourism Destinations BT. In Information and Communication Technologies
in Tourism 2014, Proceedings of the International Conference, Dublin, Ireland, 21–24 January 2014; Xiang, Z.,
Tussyadiah, I., Eds.; Springer International Publishing: Cham, Switzerland, 2013; pp. 553–564.
20.
Batty, M.; Axhausen, K.W.; Giannotti, F.; Pozdnoukhov, A.; Bazzani, A.; Wachowicz, M.; Ouzounis, G.;
Portugali, Y. Smart cities of the future. Eur. Phys. J. Spec. Top. 2012,214, 481–518. [CrossRef]
21.
Mills, J.E.; Law, R. Handbook of Consumer Behavior, Tourism and the Internet; Haworth Press Inc.: Binghamton,
NY, USA, 2005; ISBN 078902599X.
22.
Gretzel, U. Intelligent systems in tourism: A Social Science Perspective. Ann. Tour. Res.
2011
,38, 757–779.
[CrossRef]
23.
Vicini, S.; Bellini, S.; Sanna, A. How to Co-Create Internet of Things-enabled Services for Smarter Cities.
In SMART 2012: The First International Conference on Smart Systems, Devices and Technologies; IARIA: Stuttgart,
Germany, 2012; pp. 55–61.
24.
Cacho, A.; Figueredo, M.; Cassio, A.; Araujo, M.V.; Mendes, L.; Lucas, J.; Farias, H.; Coelho, J.; Cacho, N.;
Prolo, C. Social Smart Destination: A Platform to Analyze User Generated Content in Smart Tourism
Destinations. In New Advances in Information Systems and Technologies; Rocha, Á., Correia, A.M., Adeli, H.,
Reis, L.P., Mendonça Teixeira, M., Eds.; Springer International Publishing: Cham, Switzerland, 2016;
pp. 817–826. ISBN 978-3-319-31232-3.
25.
World Tourism Organization (UNWTO). Report of the First Meeting of the UNWTO Tourism Resilience Committee;
UNWTO: Madrid, Spain, 2009; pp. 1–41.
Sustainability 2017,9, 2317 17 of 19
26.
La Rocca, R.A. The Role of Tourism in Planning the Smart City. Tema. J. Land Use Mobil. Environ.
2014
,7,
269–284. [CrossRef]
27.
Boes, K.; Buhalis, D.; Inversini, A. Smart tourism destinations: Ecosystems for tourism destination
competitiveness. Int. J. Tour. Cities 2016,2, 108–124. [CrossRef]
28. Bakhtin, M.M. The Dialogic Imagination: Four Essays; University of Texas Press: Austin, TX, USA, 1983.
29.
Ooi, C.-S. Contrasting strategies: Tourism in Denmark and Singapore. Ann. Tour. Res.
2002
,29, 689–706.
[CrossRef]
30.
Gretzel, U.; Werthner, H.; Koo, C.; Lamsfus, C. Conceptual foundations for understanding smart tourism
ecosystems. Comput. Hum. Behav. 2015,50, 558–563. [CrossRef]
31.
Fistola, R.; Anna, R.; Rocca, L. Smart City Planning: A Systemic Approach. In Proceedings of the 6th
Knowledge Cities World Summit, KCWS 2013, Istanbul, Turkey, 9–12 September 2013; pp. 520–529.
32.
Al Nuaimi, E.; Al Neyadi, H.; Mohamed, N.; Al-Jaroodi, J. Applications of big data to smart cities. J. Internet
Serv. Appl. 2015,6, 25. [CrossRef]
33.
Kitchin, R. Big Data, new epistemologies and paradigm shifts. Big Data Soc.
2014
,1, 2053951714528481.
[CrossRef]
34.
Lamsfus, C.; Martín, D.; Alzua-Sorzabal, A.; Torres-Manzanera, E. Smart Tourism Destinations: An Extended
Conception of Smart Cities Focusing on Human Mobility. In Information and Communication Technologies in
Tourism 2015, Proceedings of the International Conference, Lugano, Switzerland, 3–6 February 2015; Tussyadiah, I.,
Inversini, A., Eds.; Springer International Publishing: Cham, Switzerland, 2015; pp. 363–375.
35.
Guinard, D.; Trifa, V.; Mattern, F.; Wilde, E. From the Internet of Things to the Web of Things:
Resource-oriented Architecture and Best Practices. In Architecting the Internet of Things; Uckelmann, D.,
Harrison, M., Michahelles, F., Eds.; Springer: Berlin/Heidelberg, Germany, 2011; pp. 97–129,
ISBN 978-3-642-19157-2.
36.
Komninos, N.; Pallot, M.; Schaffers, H. Special Issue on Smart Cities and the Future Internet in Europe.
J. Knowl. Econ. 2013,4, 119–134. [CrossRef]
37. Batty, M. Big data, smart cities and city planning. Dialogues Hum. Geogr. 2013,3, 274–279. [CrossRef]
38.
Yaqoob, I.; Chang, V.; Gani, A.; Mokhtar, S.; Hashem, I.A.T.; Ahmed, E.; Anuar, N.B.; Khan, S.U. Information
fusion in social big data: Foundations, state-of-the-art, applications, challenges, and future research directions.
Int. J. Inf. Manag. 2016. [CrossRef]
39.
Gani, A.; Siddiqa, A.; Shamshirband, S.; Hanum, F. A survey on indexing techniques for big data: Taxonomy
and performance evaluation. Knowl. Inf. Syst. 2016,46, 241–284. [CrossRef]
40. Chen, M.; Mao, S.; Liu, Y. Big Data: A Survey. Mob. Netw. Appl. 2014,19, 171–209. [CrossRef]
41.
Tussyadiah, I.P. An Assessment of Contagion on Social Networking Sites. In Information and Communication
Technologies in Tourism 2012, Proceedings of the International Conference, Helsingborg, Sweden, 25–27 January 2012;
Fuchs, M., Ricci, F., Cantoni, L., Eds.; Springer: Vienna, Austria, 2012; pp. 25–35.
42.
Buhalis, D.; Law, R. Progress in information technology and tourism management: 20 years on and 10 years
after the Internet—The state of eTourism research. Tour. Manag. 2008,29, 609–623. [CrossRef]
43.
Sui, D.; Goodchild, M. The convergence of GIS and social media: Challenges for GIScience. Int. J. Geogr. Inf.
Sci. 2011,25, 1737–1748. [CrossRef]
44.
Di Minin, E.; Tenkanen, H.; Toivonen, T. Prospects and challenges for social media data in conservation
science. Front. Environ. Sci. 2015,3, 63. [CrossRef]
45.
Zeng, B.; Gerritsen, R. What do we know about social media in tourism? A review. Tour. Manag. Perspect.
2014,10, 27–36. [CrossRef]
46.
Mayer-Schönberger, V.; Cukier, K. Big Data: A Revolution that Will Transform How We Live, Work, and Think;
An Eamon Dolan Book; Houghton Mifflin Harcourt: Boston, MA, USA, 2013; ISBN 9780544002692.
47.
Arts, K.; van der Wal, R.; Adams, W.M. Digital technology and the conservation of nature. Ambio
2015
,44,
661–673. [CrossRef] [PubMed]
48.
Hausmann, A.; Toivonen, T.; Slotow, R.; Tenkanen, H.; Moilanen, A.; Heikinheimo, V.; Di Minin, E. Social
Media Data Can Be Used to Understand Tourists’ Preferences for Nature-Based Experiences in Protected
Areas. Conserv. Lett. 2017. [CrossRef]
49.
Tufekci, Z. Big Questions for social media big data: Representativeness, validity and other methodological
pitfalls. In Proceedings of the 8th International Conference on Weblogs and Social Media, ICWSM 2014,
Ann Arbor, MI, USA, 1–4 June 2014; The AAAI Press: Ann Arbor, MI, USA, 2014; pp. 505–514.
Sustainability 2017,9, 2317 18 of 19
50.
MacKay, K.; Vogt, C. Information technology in everyday and vacation contexts. Ann. Tour. Res.
2012
,39,
1380–1401. [CrossRef]
51.
Höpken, W.; Fuchs, M.; Zanker, M.; Beer, T. Context-Based Adaptation of Mobile Applications in Tourism.
Inf. Technol. Tour. 2010,12, 175–195. [CrossRef]
52.
Fuchs, M.; Höpken, W.; Lexhagen, M. Big data analytics for knowledge generation in tourism destinations—
A case from Sweden. J. Destin. Mark. Manag. 2014,3, 198–209. [CrossRef]
53.
Höpken, W.; Fuchs, M.; Lexhagen, M. The knowledge destination—Applying methods of business
intelligence to tourism. In Encyclopedia of Business Analytics and Optimization; IGI Glob: Hershey, PA,
USA, 2014; pp. 307–321.
54.
Fuchs, M.; Eybl, A.; Höpken, W. Successfully selling accommodation packages at online auctions—The case
of eBay Austria. Tour. Manag. 2011,32, 1166–1175. [CrossRef]
55.
Girardin, F.; Fiore, F.D.; Ratti, C.; Blat, J. Leveraging explicitly disclosed location information to understand
tourist dynamics: A case study. J. Locat. Based Serv. 2008,2, 41–56. [CrossRef]
56.
Kádár, B. Measuring tourist activities in cities using geotagged photography. Tour. Geogr.
2014
,16, 88–104.
[CrossRef]
57.
García-Palomares, J.C.; Gutiérrez, J.; Mínguez, C. Identification of tourist hot spots based on social networks:
A comparative analysis of European metropolises using photo-sharing services and GIS. Appl. Geogr.
2015
,
63, 408–417. [CrossRef]
58.
Instituto Nacional de Estatística. Censos 2011 Resultados Definitivos—Região Lisboa; Instituto Nacional de
Estatística: Lisboa, Portugal, 2012; ISBN 978-989-25-0185-7.
59.
Entidade Regional de Turismo da Região de LIsboa (ERT-RL). Plano Estratégico para o Turismo na Região
de Lisboa 2015–2019. Available online: http://www.ertlisboa.pt/fotos/editor2/lis_9970_04248_007_15.pdf
(accessed on 11 December 2017).
60.
Instituto Nacional de Estatística. Estatísticas do Turismo 2016; Instituto Nacional de Estatística: Lisboa,
Portugal, 2017; ISBN 978-989-25-0396-7.
61.
Instituto Nacional de Estatística. Estatísticas do Turismo 2014; Instituto Nacional de Estatística: Lisboa,
Portugal, 2015; ISBN 978-989-25-0308-0.
62.
Observatório de Turismo de Lisboa. Survey to the Purpose of Trip in Lisbon City 2014. Available
online: https://www.visitlisboa.com/sites/default/files/2016-10/Survey%20on%20the%20Purpose%
20of%20Trip%202014%20-%20Lisbon%20City.pdf (accessed on 11 December 2017).
63.
Observatório Turismo de Lisboa. Visitors Activities and Information Survey in Lisbon Region 2015. Available
online: https://www.visitlisboa.com/sites/default/files/2016-10/VISITOR%20ACTIVITIES%20AND%
20INFORMATION%20SURVEY%202015.pdf (accessed on 11 December 2017).
64. Anselin, L. Local Indicators of Spatial Association—LISA. Geogr. Anal. 1995,27, 93–115. [CrossRef]
65. Berry, W.; Feldman, S. Multiple Regression in Practice; Sage: Thousand Oaks, CA, USA, 1985.
66. Hair, J.F. Multivariate Data Analysis with Readings; Prentice Hall: Upper Saddle River, NY, USA, 1995.
67.
Ringle, C.; Wende, S.; Becker, J.-M. SmartPLS 3. Available online: https://www.smartpls.com/ (accessed on
11 December 2017).
68.
Salvati, L.; Gargiulo Morelli, V.; Weijnen, M.; van Bueren, E.; Wenzler, I.; De Reuver, M. Towards
Intelligently—Sustainable Cities? Tema. J. Land Use Mobil. Environ. 2013,6, 73–86.
69. Andrejevic, M.; Burdon, M. Defining the Sensor Society. Telev. New Media 2014,16, 19–36. [CrossRef]
70.
Del Chiappa, G.; Baggio, R. Knowledge transfer in smart tourism destinations: Analyzing the effects of a
network structure. J. Destin. Mark. Manag. 2015,4, 145–150. [CrossRef]
71.
Chen, L.; Ng, E.; Huang, S.-C.; Fang, W.-T. A Self-Evaluation System of Quality Planning for Tourist
Attractions in Taiwan: An Integrated AHP-Delphi Approach from Career Professionals. Sustainability
2017
,
9, 1751. [CrossRef]
72.
Paul, P. Tallon Corporate Governance of Big Data: Perspectives on Value, Risk, and Cost. Computer
2013
,46,
32–38. [CrossRef]
73. Minghetti, V.; Buhalis, D. Digital Divide in Tourism. J. Travel Res. 2009,49, 267–281. [CrossRef]
74.
Tachizawa, E.M.; Alvarez-Gil, M.J.; Montes-Sancho, M.J. How “smart cities” will change supply chain
management. Supply Chain Manag. Int. J. 2015,20, 237–248. [CrossRef]
Sustainability 2017,9, 2317 19 of 19
75.
Saravanan, S.; Sadhu Ramakrishnan, B. Preserving privacy in the context of location based services through
location hider in mobile-tourism. Inf. Technol. Tour. 2016,16, 229–248. [CrossRef]
76. Bruns, K.; Jacob, F. Value-in-Use and Mobile Technologies. Bus. Inf. Syst. Eng. 2014,6, 349–359. [CrossRef]
©
2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).