Journal of Civil Engineering and Urbanism  
Volume 10, Issue 4: 35-41; July 25, 2020  
Spatio-Temporal Understanding and Representation of Transformative  
Urban Mobility and Trip Patterns, A Review  
Arian Behradfar1 and Soheil Mohammadi2Error! Bookmark not defined.  
1Department of Geomatics and Spatial Information Engineering, College of Engineering, University of Tehran, Tehran 1439957131, Iran  
2Department of Naval, Electrical, Electronic and Telecommunications Engineering, DITEN, University of Genoa, Genoa 16126, Italy  
Corresponding author’s Email:  
The rapid development in transport system monitoring provide planners and researchers with new opportunities to  
understand the trends of mobility patterns in urban areas, known as transformation of urban mobility. It brings  
businesses and cities together to implement system-level integrated initiatives to conducting urban mobility and  
transport system toward a more efficient future. As a result, identification of the trip patterns and spatio-temporal  
dependencies in urban areas requires a comprehensive understanding of high-dimensional human mobility  
dynamics. These emerging trends need a framework to identify urban mobility patterns from a spatio-temporal  
perspective that includes various visualized representation of mobility patterns and travel behaviour. The main  
purpose of this study is to investigate different data sources and methods used in the literature to obtain the  
proposed patterns in urban areas. The spatio-temporal models evaluated in this review can be used in a wide range  
of mobility studies suggesting trip patterns and related variables are significantly affected by spatial and non-spatial  
Keywords: Transformation of urban mobility, Mobility flows, Trip patterns, spatio-temporal dependencies, spatial  
creation and deployment of sustainability. As a result, the  
identification of main individual aspects that provide vital  
information on the progress and effectiveness of mobility  
transformation is completely essential in transportation  
engineering. These individual aspects have to comprise  
various characteristics of interaction trends and precisely  
express mobility specifics in urban areas so that  
simultaneously take into account the three dimensions of  
sustainability: social, economic and environmental issues  
Mobility can be considered as an important socio-  
economic resource and articulator in society, since it is  
directly related to the movement of people between  
different spatio-temporal hierarchies. The rapid growth in  
urban areas, private vehicle fleet and the lack of proper  
planning of transportation systems have led to increasing  
deterioration of mobility, emerging different social  
challenges and environmental problems (Costa et al.,  
One of the best sources for monitoring the  
transformation of urban mobility is obtaining daily  
mobility patterns of an urban area. These regular mobility  
flows are reflected in different aspects of people’s  
movement trajectory and transport system structure. Urban  
trips distribution, travel behavior, spatial interaction,  
activity identification, temporal dependencies, and trip  
pattern segments. This information can lead to  
understanding the current structure of urban mobility and  
existing trends and challenges in large-scale from a spatio-  
temporal perspective (Sun et al., 2016).  
Urban areas will require mobility solutions and  
changes that are sustainable, affordable, inclusive and  
integrated with people-centric infrastructure and services.  
These solutions rest on the transformation of urban  
mobility that includes a holistic and systemic approach  
that acts in the intersection between mobility  
infrastructure, social benefits, economic efficiency, and  
environmental impact. The transformation of urban  
mobility delivers a cohesive set of initiatives that act at the  
system level. Each co-designed action aims to remove  
barriers and mobilize stakeholders to contribute to  
transforming mobility (Hegyi et al., 2019).  
As we discussed, the main purpose of this study is to  
understand the trip patterns in urban areas based on travel  
behaviour analysis. Deriving the major characteristics of  
The future of mobility represents opportunities for  
cities and planners to collaborate and accelerate the  
To cite this paper: Behradfar A and Mohammadi S. (2020). Spatio-Temporal Understanding and Representation of Transformative Urban Mobility and Trip Patterns, A Review. J.  
Behradfar and Mohammadi, 2020  
mobility system and the necessitate trends in different  
to return to a few highly frequented locations (Gonzalez et  
al., 2008). The diversity of spatio-temporal regularity was  
found to be constrained, providing an encouraging  
foundation for studying and analysing mobility (Song et  
parts of the mobility strategies makes it possible to analyse  
different aspects of the transformation of urban mobility.  
The emergence of such trends with high-order interactions  
among space, time and the mentioned individual attributes  
indeed brings us new opportunities to integrate more  
knowledge into urban planning and decision-making  
system in urban areas.  
In this regard and as a literature review, different  
location-based data sources and proposed spatial analysis  
have been taken into account in the identification of urban  
mobility and trip patterns have been considered and  
discussed. Furthermore, the most frequent models and  
methods in the literature will be introduced and various  
spatial and non-spatial aspects have been taken into  
Another emerging area of urban mobility studies  
relies on network theory and statistical tools. These studies  
that termed as computational science focus on analysing  
social interactions and temporal attributes of mobility  
patterns (Hamedmoghadam et al., 2019). In a series of  
papers, GIS-based accessibility modelling and network  
analysis have been used to trip generation assessment and  
patterns of trip occurrence in accessibility infrastructure  
(Castanho et al., 2020), human mobility patterns with the  
support of location-based services (Ebrahimpour et al.,  
2020: Habibi et al, 2020), similarity and dissimilarity of  
the dynamic mobility patterns on the basis on spatio-  
temporal characteristics (Yuan et al., 2012).  
Location tracking and location-based service patterns  
are widely being used in a growing number of applications  
2015). To contribute to the transformation of urban  
mobility, studies have focus on geographical area  
subdivision-based mobility patterns with distinct  
characteristics. Different clustering algorithms and feature  
extraction methods have been used to detect spatial  
operation patterns in urban mobility (Kang et al., 2016).  
Mobility behaviour analysis, trip distribution, activity  
identification, understanding spatio-temporal interactions  
and social characteristics are the most common ways that  
represents reliable sources to analyze mobility patterns in  
Various data mining methods have been developed  
to uncover transit behaviour patterns on the basis of  
heterogeneous geospatial datasets, including public  
transport-oriented and passenger-oriented approaches for  
mobility analysis (El Mahrsi et al., 2017), spatial affinity  
propagation with spatial-behavioural features in trip  
segments (Kieu et al., 2018), trip chaining methods for  
pattern estimation, validation and diversification (Li et al.,  
2018), and choice modelling for activity identification and  
The first efforts to learn human mobility patterns were  
associated with classic transport sciences. Since the  
nineteenth century and as time-use or time-budget studies,  
various measures have been taken to identify different  
activities during the day (Szalai, 1966). Transportation  
forecasting was one of the first research fields that made  
use of mobility analysis, and mobility patterns have been  
studied since the late 1950s. The first analytical models  
implemented pre-calculated probability distributions of  
possible urban trip patterns mainly based on land use and  
socio-demographic characteristics (Douglass et al., 1957).  
Starting in the 2000s, this approach was gradually  
enriched by activity-based models, e.g., computerized  
models that use agent-based simulations, where the agents  
are modelled and driven by specific human activity such  
as work, leisure, and shopping (Kitamura et al., 2000).  
Over recent years and with the fast development of  
information and communication technologies (ICT) and  
intelligent transportation system (ITS), large quantities of  
digital traces and geo-referenced social media data plus  
transit smart cards and time table data that register  
individual activity and trip behaviour at both spatial and  
temporal scales have become available (Batty et al., 2012).  
These datasets enable planners and researchers to model  
and analyse spatio-temporal interactions in urban mobility  
based on variety of proxies on human activities (Han et al.,  
Over the past few years, many studies have been  
conducted to explore urban trip patterns using various  
modelling and analytic approaches based on massive  
human mobility data, such as optimization-based routing  
equilibrium models for congestion alleviation (Çolak et  
al., 2016), clustering-based correlated analyses of mobility  
similarities and relationships, low-level mobility pattern  
discovery and multi-scale exploration of social  
Large-scale studies showed human trajectories have  
a high degree of temporal and spatial regularity, each  
individual being characterized by a time-independent  
characteristic travel distance and a significant probability  
J. Civil Eng. Urban., 10 (4): 35-41, 2020  
fragmentation (Schneider et al., 2013). With the  
activity identification along with priorities, types and  
scheduling. In fact, travel behavior is a part of complex  
hierarchical structures that include activity trends, mobility  
flows, spatial interactions, and finally, activity  
identification to derive mobility patterns in urban areas.  
In general, the total number of daily stops,  
particularly between relatively distinct units constructs  
urban trips. Therefore, travel behavior can be defined as  
the spatial way sequenced segments of the urban trip take  
place. There have been considerable efforts extended to  
understanding the spatio-temporal characteristics and  
complexity of travel behavior using a variety of  
abstraction level in modelling (Pritchard et al., 2014).  
availability of massive human mobility data, machine  
learning techniques have been playing a more and more  
important role in gaining a deep understanding of human  
mobility behaviour (Toch et al., 2019), ranging from  
movement pattern mining (Chen et al., 2016), mobility  
prediction (Ouyang et al., 2018), and movement mode  
classification to lifestyle discovering and prediction (Chen  
Travel behaviour analysis toward transformative  
urban mobility  
Starting from the premise that the mobility is a  
fundamental issue, discovering the transformation of urban  
mobility impact on different parts of the urban area is  
essential. Mobility paradigm shifting in cities happens in a  
wide range of other contexts that are host of problems and  
challenges including urban accessibility degradation, lack  
of social equity, land use conflicts, urban sprawl, and  
social exclusion (Bibri, 2019: Naghdi et al, 2016). Given  
these issues, the transformation of urban mobility  
appraisal guidance includes a spectrum of various impacts  
through qualitative and quantitative assessments and  
Visualized representation of mobility patterns  
Temporal distribution of urban mobility  
The high resolution temporal distribution of boarding  
times during a 24-hour period includes the number of trips  
per minute for any moment. The main advantage of this  
diagram is the explicit representation of urban trip peaks in  
a day. The diagram in figure 1 indicates that the sharp  
peaks in urban mobility occur in starting (8 am) and  
ending (7pm) hours of working time in Seol, South Korea  
In-depth knowledge of spatio-temporal distribution  
of urban mobility is a key aspect in characterizing the  
behaviour of transit users in urban areas for mobility  
transforming identification. This information can provide  
valuable insight into urban trips analysis such as when  
they begin and end within complete day. Spatial  
observations of urban trip distribution can indicate  
infrastructure efficiency, transport accessibility and  
modality, and traffic congestions in urban areas. Mode  
share trends based on daily time, weather condition,  
season, and other features allow transit planners to remark  
mobility transformation to promote their services (Sun et  
Since urban mobility trends are based on the  
particular purposes, utilizing integrated attributes like the  
trip duration, type and regularity can lead to assigning  
activities to individual stakeholders for the activity-based  
model. Home, work or educational units can be considered  
along with socio-demographic attributes for activity-based  
micro-simulations (Gan et al., 2020).  
The proposed discussion starts with observations  
concerning the relationship between the spatial and non-  
spatial conceptualization of travel behavior based on  
Figure 1. High temporal resolution of trip distribution (Ali  
Accurate distribution of mobility patterns based on  
temporal features along with public transport schedule,  
private vehicles flows, travel time tables, and qualifying  
transfer volume by different modes can lead to a large  
scale agent-based simulation model of urban mobility.  
Behradfar and Mohammadi, 2020  
Advanced statistical approaches for travel behavior  
Spatio-temporal flows of mobility  
analysis and activity identification provide a new insight  
of mobility patterns to transit planning. Furthermore, it can  
be considered as an effective framework to monitor the  
transformation of urban mobility and transport  
infrastructure performance.  
As shown in figure 3, the total mobility flows  
entering a distinct region are inflow and the ones leaving a  
region for another denote outflow. Both flows track the  
mobility patterns between different regions of urban area.  
these flows are very applicable in risk assessment, transport  
management and particularly, the transformation of urban  
mobility studies. They can accurately reflect the spatio-  
temporal dependencies of urban mobility such as spatial  
correlation, traffic congestion and surrounding conditions.  
Furthermore, it is possible to employ the spatio-  
temporal residual network analysis methods to model  
nearby and distant spatial dependencies of any two regions,  
temporal closeness, period and trends, dynamical  
aggregated outputs, and assigned weights. In this way, the  
effects of different components, internal structures and  
external factors on proposed flows can be evaluated (Zhang  
Spatial distribution of urban mobility  
The trip patterns extracted from large-scale datasets  
at spatially-aggregated levels, as shown in figure 2, can  
efficiently contribute to travel behavior analysis in urban  
areas. These visualized representations of the spatial  
distribution of urban mobility cover a wide range of  
impacts, variables, flows, paradigm shifts, and attributes  
related to transportation network in urban areas (Bao et  
all., 2018). These GIS-based models can be used in  
mobility patterns recognition, travel behavior, activity  
identification, land use planning, and socio-economic  
impacts in urban areas. They also are the core of spatially  
integrated approaches for sustainability. The inclusion of  
urban trip patterns extracted from such spatial analysis  
significantly improve the performance of the transport  
system and logistic infrastructure. They can be an  
aggregated basis of different quantitative and qualitative  
comparison analysis for various purposes (Haselsteiner et  
Figure 3. Regional mobility inflow and outflow (a) and  
measurements (b) - (Zhang et al., 2017).  
Figure 2. Spatial distribution of the coefficient of urban  
trip patterns variables (Bao et all., 2018).  
J. Civil Eng. Urban., 10 (4): 35-41, 2020  
Spatial interaction estimation  
The emergence of different mobility monitoring  
systems brings new opportunities to integrate complex and  
higher-order interactions among space, time and attributes.  
Analytical frameworks and statistical techniques provide  
better understating and representation ways to better  
interpret mobility patterns and urban dynamics. In this  
regard, by using factorization model to decompose high-  
dimensional mobility data into specific patterns, from  
which we can extract key information by reasoning about  
the semantics of regions and activities in urban areas (Sun  
Once the urban trips are identified, additional  
attributes such as zone properties, coordinates, distances,  
and travel mode are considered to generate origin-  
destination matrixes (OD) for zone analysis. OD matrix  
integrates activity identification data to assign activities to  
individual users and trips for the activity-based model, as  
different spatial interactions, shown in figure 4. Spatial  
interaction models illustrate how different zones are  
functionally interdependent. They can also reflect the  
human-land relationship has long been a core topic in  
urban planning and transport studies, as well as the  
transformation of urban mobility.  
The present study investigated how the trip patterns and  
related spatial characteristics and variables can be  
extracted from different data sources to contribute to the  
transformation of urban mobility. Trip pattern information  
was estimated with the trip generation models developed  
based on electronic fare payment system data, transit smart  
card data, geo-referenced social media data, large scale  
global positioning system (GPS), trip flows based on the  
public transport system and scheduled time table, and road  
network attributes. Understanding the urban trip flows  
mobility patterns and their trends and paradigm shifts is an  
immense step toward sustainability in spatial planning.  
Geospatial Information System (GIS) plays an undeniable  
role in such field. With the advantages of spatio-temporal  
analysis, it is possible to evaluate the urban mobility  
dynamics based on trip behaviour, distribution of trips  
based on various conditions, activity identification, and  
departure time table of different public transport modes.  
As the transformation of urban mobility has direct effects  
on different aspect of society, understanding these changes  
can accelerate the creation and deployment of sustainable  
mobility and transport system in urban areas.  
With the advances in computer science and  
technology, the Intelligent Transport Systems (ITS) are  
emerging as one of the best solution to tackle the transport  
system by providing accurate and real-time data.  
Advanced data mining techniques are implemented to  
derive the travel behavior patterns and characteristics to  
evaluate the unexpected trends in mobility dynamics.  
The reviewed models extracted from proposed data  
sources by mean of GIS focus on providing a generalized  
data-driven framework to better utilize the increasing  
amount of individual-based mobility trends, underlying  
spatio-temporal structure of urban areas.  
The questions regarding mobility patterns and  
required action as a response to trends, dynamics and  
complexity of urban areas are essential to  
spatial/transportation planning. Our review enriches the  
information in mobility data by discussing trip patterns in  
a multi-dimensional setting, accurate mapping and data  
analysis provided by GIS. Comprehensive understanding  
of spatio-temporal urban dynamics through collective  
transit mobility demonstrates great flexibility in studying  
several directions of transformative urban mobility such as  
trends in trip behavior, trip purposes, trip chain and  
transport modes. Proposed aspects of mobility in urban  
areas can be inferred and integrated to the GIS-based  
models and analysis to add a new dimension to the  
Figure 4. Spatial interactions based on distinct origin-  
destination for spatio-temporal patterns (Sun et al., 2016).  
Behradfar and Mohammadi, 2020  
Using Large-Scale Social Media Data. ISPRS International Journal  
of Geo-Information. Feb;9(2):125. Google Scholar,  
planning process. In this way, we can use spatial  
interaction, trip distribution and dependencies as priors to  
registering individual trip behavior analysis, offering new  
insights in mobility studies by enhancing transport system  
and infrastructure in urban areas.  
Gan Z, Yang M, Feng T, Timmermans H. (2020). Understanding urban  
mobility patterns from a spatiotemporal perspective: daily ridership  
profiles of metro stations. Transportation. Feb;47(1):315-36. Google  
Gheitarani N, El-Sayed S, Cloutier S, Budruk M, Gibbons L, Khanian M.  
(2020). Investigating the Mechanism of Place and Community  
Impact on Quality of Life of Rural-Urban Migrants. International  
Journal of Community Well-Being. Mar;3(1):21-38. Google Scholar  
Author’s contribution  
Both authors contributed equally to this work.  
Gonzalez MC, Hidalgo CA, Barabasi AL. (2008). Understanding  
individual human mobility patterns. nature. Jun;453(7196):779-82.  
Competing interests  
The authors declare that they has no competing  
Habibi K, Hoseini SM, Dehshti M, Khanian M, Mosavi A. (2020). The  
Impact of Natural Elements on Environmental Comfort in the  
Iranian-Islamic Historical City of Isfahan. International Journal of  
Environmental Research and Public Health. Jan;17(16):5776.  
Hamedmoghadam H, Ramezani M, Saberi M. (2019). Revealing latent  
characteristics of mobility networks with coarse-graining. Scientific  
reports. May 17;9(1):1-0. Google Scholar  
Han G, Sohn K. (2016). Activity imputation for trip-chains elicited from  
Ali A, Kim J, Lee S. (2016). Travel behavior analysis using smart card  
data. KSCE Journal of Civil Engineering. May 1;20(4):1532-9.  
smart-card data using  
a continuous hidden Markov model.  
Transportation Research Part B: Methodological. Jan 1;83:121-35.  
Bao J, Liu P, Qin X, Zhou H. (2018). Understanding the effects of trip  
patterns on spatially aggregated crashes with large-scale taxi GPS  
data. Accident Analysis & Prevention. Nov 1;120:281-94. Google  
Haselsteiner E, Smetschka B, Remesch A, Gaube V. (2015). Time-use  
patterns and sustainable urban form: A case study to explore  
potential links. Sustainability. Jun;7(6):8022-50.Google Scholar,  
Batty M, Axhausen KW, Giannotti F, Pozdnoukhov A, Bazzani A,  
Wachowicz M, Ouzounis G, Portugali Y. (2012). Smart cities of the  
future. The European Physical Journal Special Topics. Nov  
1;214(1):481-518. Google Scholar  
Hegyi FB, Morgen H, Vendel M. (2019). Urban mobility in  
transformation: demands on education to close the predicted  
knowledge gap. Transforming Urban Mobility. Oct:114. Google  
Bibri SE. (2019). On the sustainability of smart and smarter cities in the  
era of big data: an interdisciplinary and transdisciplinary literature  
review. Journal of Big Data. Dec 1;6(1):25.Google Scholar  
Kahvand M, Gheitarani N, Khanian MO, Ghadarjani RA. (2015). Urban  
solid waste landfill selection by SDSS. Case study: Hamadan.  
Environment Protection Engineering. 41(2).Google Scholar, doi  
Camagni R, Gibelli MC, Rigamonti P. (2002). Urban mobility and urban  
form: the social and environmental costs of different patterns of  
urban expansion. Ecological economics. Feb 1;40(2):199-216.  
Kang C, Qin K. (2016). Understanding operation behaviors of taxicabs in  
cities by matrix factorization. Computers, Environment and Urban  
Castanho RA, Behradfar A, Vulevic A, Naranjo Gómez JM. (2020)  
Analyzing Transportation Sustainability in the Canary Islands  
Archipelago. Infrastructures. Jul;5(7):58. Google Scholar,  
Khanian M, Marshall N, Zakerhaghighi K, Salimi M, Naghdi A. (2018).  
Transforming agriculture to climate change in Famenin County,  
West Iran through a focus on environmental, economic and social  
factors. Weather and climate extremes. Sep 1;21:52-64.Google  
Chen R, Chen M, Li W, Wang J, Yao X. (2019). Mobility modes  
awareness from trajectories based on clustering and a convolutional  
neural network. ISPRS International Journal of Geo-Information.  
Kieu, L, M. Ou, Y, Cai, C. (2018). Large-scale transit market  
segmentation with spatial-behavioural features. Transportation  
Research Part C: Emerging Technologies, 90, 97-113. Google  
Chen X, Shi D, Zhao B, Liu F. (2016). Mining individual mobility  
patterns based on location history. In2016 IEEE First International  
Conference on Data Science in Cyberspace (DSC) Jun 13 (pp. 252-  
Kitamura R, Chen C, Pendyala RM, Narayanan R. (2000). Micro-  
simulation of daily activity-travel patterns for travel demand  
forecasting. Transportation. Feb 1;27(1):25-51. Google Scholar  
Çolak S, Lima A, González MC. (2016). Understanding congested travel  
in urban areas. Nature communications. Mar 15;7(1):1-8. Google  
Li T, Sun D, Jing P, Yang K. (2018). Smart card data mining of public  
transport destination:  
literature review. Information.  
Costa PB, Neto GM, Bertolde AI. (2017). Urban mobility indexes: A  
brief review of the literature. Transportation research procedia. Jan  
Naghdi A, Khanian M, Rueentan M. (2016). The urban dilemmas in Iran  
marginal urban area; A case study of Kermanshah city. Journal of  
Civil Engineering and Urbanism.1:16-23. Google Scholar, PDF  
Douglass Carroll JD, Bovls HW. (1957). Predicting local travel in urban  
regions. Papers in Regional Science. 3(1):183-97.Google Scholar,  
Nielsen TA, Næss P. (2012). Activity based indicators of connections  
and access needs (ABICA). InAccessibility Instruments for Planning  
Practice: COST Action TU1002 (pp. 71-76). COST Office. Google  
Ebrahimpour Z, Wan W, Velázquez García JL, Cervantes O, Hou L.  
(2020). Analyzing Social-Geographic Human Mobility Patterns  
J. Civil Eng. Urban., 10 (4): 35-41, 2020  
Ouyang X, Zhang C, Zhou P, Jiang H, Gong S. (2016). Deepspace: An  
The Royal Society Interface. 2015 Jan 6;12(102): 1089. Google  
online deep learning framework for mobile big data to understand  
human mobility patterns. arXiv preprint arXiv:1610.07009. Oct 22.  
Szalai A. (1966). The multinational comparative time budget research  
project: a venture in international research cooperation.Google  
Pritchard JP, Moura F, e Silva JD, Martinez LM. (2014). Spatial analysis  
of transportation-related social exclusion in the Lisbon metropolitan  
area. Procedia-Social and Behavioral Sciences. Feb 5;111:440-9.  
Toch E, Lerner B, Ben-Zion E, Ben-Gal I. (2019). Analyzing large-scale  
human mobility data: a survey of machine learning methods and  
applications. Knowledge and Information Systems. Mar 5;58(3):501-  
Sarvar H, Amini J, Laleh-Poor M. (2011). Assessment of risk caused by  
earthquake in region 1 of Tehran using the combination of RADIUS,  
TOPSIS and AHP models. Journal of Civil Engineering and  
Urbanism.;1(1):39-48. Google Scholar, PDF  
Wang Y, de Almeida Correia GH, de Romph E, Timmermans HJ. (2017).  
Using metro smart card data to model location choice of after-work  
activities: An application to Shanghai. Journal of Transport  
Schneider CM, Rudloff C, Bauer D, González MC. (2013). Daily travel  
behavior: lessons from a week-long survey for the extraction of  
human mobility motifs related information. InProceedings of the 2nd  
ACM SIGKDD international workshop on urban computing Aug 11  
Yuan Y, Raubal M. (2012). Extracting dynamic urban mobility patterns  
from mobile phone data. InInternational conference on geographic  
information scienceSep 18 (pp. 354-367). Springer, Berlin,  
Song C, Qu Z, Blumm N, Barabási AL. (2010). Limits of predictability  
in human mobility. Science. Feb 19;327(5968):1018-21. Google  
Scholar, DOI: 10.1126/science.1177170  
Zaker Haghighi K, Gheitarani N, Khanian M, Taghadosi R. (2014).  
Examination of effects of urban street configuration on the amount  
of commercial buildings establishment (according to natural  
movement theory), Case study: Hamedan. European Online Journal  
of Natural and Social Sciences. Jan 23;3(1):pp-20.. Google Scholar  
Sun L, Axhausen KW. (2016). Understanding urban mobility patterns  
with a probabilistic tensor factorization framework. Transportation  
Research Part B: Methodological. Sep 1;91:511-24. Google Scholar,  
Zhang J, Zheng Y, Qi D. (2016). Deep spatio-temporal residual networks  
for citywide crowd flows prediction. arXiv preprint  
arXiv:1610.00081. Oct 1.Google Scholar  
Sun L, Jin JG, Axhausen KW, Lee DH, Cebrian M. (2014). Quantifying  
long-term evolution of intra-urban spatial interactions. Journal of