Content uploaded by Harvey Miller
Author content
All content in this area was uploaded by Harvey Miller on Mar 02, 2016
Content may be subject to copyright.
Geography Compass 9/4 (2015): 180–189, 10.1111/gec3.12204
Geographic Information Systems for Transportation in the
21st Century
Harvey J. Miller
1
*
and Shih-Lung Shaw
2
1
Department of Geography, The Ohio State University
2
Department of Geography, University of Tennessee
Abstract
Geographic Information Systems for Transportation: Principles and Applications (Miller and Shaw 2001)
remains the only major authored text on the interdisciplinary field of Geographic Information Systems for
Transportation (GIS-T). However, Miller and Shaw (2001) is a product of the 20th century, and the
fields of GIS, transportation and GIS-T have changed dramatically in the early 21st century. We are
witnessing a revolution in transportation and urban sciences, fueled by a stunning advancement in capa-
bilities to capture, store and process data, as well as communicate information and knowledge derived
from these data. This paper is a review of GIS-T in the 20th and 21st centuries. Using Miller and Shaw
(2001) as a touchstone, we discuss elements of GIS-T that have stood the test of time as well as new
technologies and ideas that an updated GIS-T canon should include.
1. Introduction
Geographic Information Systems for Transportation: Principles and Applications (Miller and Shaw 2001)
remains the only major authored text on the interdisciplinary field of Geographic Information
Systems for Transportation (GIS-T). Although published in the second year of the 21st century,
Miller and Shaw (2001) is a product of the 20th century, as its bibliography suggests. It is roughly
15 years since it was published, but the fields of GIS, transportation and GIS-T have changed
dramatically.
It is not hyperbole to state that we are witnessing a revolution in the human sciences, includ-
ing transportation and urban sciences, fueled by a stunning advancement in capabilities to
capture, store and process data, as well as communicate information and knowledge derived
from these data. These greatly expanded capabilities have spawned related technologies and
scientific fields. These include location-based services (LBS): the delivery of information based
on geographic context in real time. It also includes the interdisciplinary fields of mobility science
(data-driven modeling of human and animal dynamics using location data) and computational
transportation science (the application of computer science to understand and manage
transportation systems) (Winter et al. 2010).
This paper is a review of GIS-T in the 20th and 21st centuries. Using Miller and Shaw (2001) as
a touchstone, we discuss elements of GIS-T that have stood the test of time as well as new tech-
nologies and ideas that an updated GIS-T canon should include. The next section of this paper
reviews GIS-T in the 20th century: these are the enduring principles that still should be taught
by professors and understood by professionals. Section 3 discusses GIS-T in the 21st century: tech-
nologies and issues that have emerged from the mobility revolution and seem likely to persist. The
final section concludes the paper with some brief remarks about GIS-T educational challenges.
2. Enduring Principles from 20th Century GIS-T
The big change in contemporary GIS-T relative to the late 20th century is the shift from a data-
poor and computation-poor environment to one that is abundant, perhaps even overwhelming,
© 2015 The Author(s)
Geography Compass © 2015 John Wiley & Sons Ltd
in both. While this has been developing for a long time, the exponential growth described by
Moore’s Law (superlinear growth in computing speed with a doubling period of 18months)
implies cumulative changes leading to a fundamental shift in how we approach science, man-
agement and our lives. This is not to say all of the data we desire is at our fingertips or that all
problems can be solved through computation. But anyone who has worked in GIS-T for the
past two decades will attest that we are in a new era.
What is enduring from the GIS-T canon of the 20th century? A more dramatic way to ask
this question – can we now say that GIS-T is not just a bundle of technologies with the same
application but rather a science with basic principles (Goodchild 1992)? We believe that the
answer is a confident “yes.” Despite the revolution, a scientist or planner working in GIS-T still
must understand how transportation and other geographic data are referenced to the earth, care-
fully manage transportation databases, know how to process networks to solve fundamental
routing problems and appropriately use spatial analytical and geovisualization techniques. This
section reviews these enduring principles and their progress in the 21st century.
2.1. TRANSPORTATION DATA MODELING
The heart of any GIS-T project is a georeferenced transportation database. Most likely, this da-
tabase will be focused on a spatial network. A spatial network is a logical structure representing
locations and connections that exist within a spatial framework where distance and direction
are meaningful. In our case, the framework is a digital representation of the earth’s surface
and possibly related locations above or below the surface (e.g., subway lines and
airline routes). With transportation systems, nodes correspond to intersections, junctions, stops
or switching points, and arcs represent infrastructure and/or services between nodes. Spatial
networks are sparse: each node is only connected to a small number of the other nodes in the
network. These are typically the nearest neighbors of each node (e.g., street networks) but
not always (e.g., airline networks). Nevertheless, since these networks in the real world are
constrained by physics and geography, there will be some spatial regularity to these connections,
shaped by phenomena such as terrain, urban hierarchies and political boundaries.
GIS-T is worthy of research and careers since georeferencing transportation network data
creates new applications that were not possible when we handled networks as embedded in
space but devoid of geography. Georeferencing allows accurate representation and intuitive
visualization and also allows other relevant economic, social, biological and physical data to
be integrated.
The centrality of georeferenced spatial network data means that core GIS-T knowledge
includes mapping, graph theory and database design. Basic mapping concepts include geodetic
datums (models of the earth’s shape), methods for projecting data from a datum to a plane,
coordinate systems for referencing locations within that plane and the concepts of map scale
and spatial error. Graph theory is the basic mathematics underlying network representation.
Connectivity, directed arcs, paths, routes, trees and their mathematical representation are
essential to the fundamental problem in transportation science: finding shortest paths within
networks (more on this below). Database design involves techniques for modeling data at the
conceptual (what the data means), logical (how the user will interact with the data) and physical
(how the data will be stored and accessed in the computer) levels.
In 20th century GIS-T, there were two major paths for data modeling. One was as-
sociated with relational database technologies. In relational databases, the user interacts
with the data represented as tables, constructed and linked in precise ways. Entity-
relationship models support conceptual data modeling in a manner that can be translated
to an initial relational schema ( design) and refined to a simpler form through the
GIS-T in the 21st Century 181
© 2015 The Author(s)
Geography Compass © 2015 John Wiley & Sons Ltd
Geography Compass 9/4 (2015): 180–189, 10.1111/gec3.12204
normalization process. One could also build aggregate l inear f eat ure s such as named or
numbered routes through a street and highway network, as well as easily reference data
such as street addresses. This works well for transportation networks. However, it is
cumbersome for representing geometry and therefore geography. It also is not well-
suited for data that changes its condition frequently, such as the real-time location
and status of people and vehicles within a transportation system. I t is also a strange
way of thinking about real-world entities – as bundles of attributes rather than things.
Object-orientation (OO), t he second path, is a more natural way of thinking about entities
in the real world (interacting objects with behaviors). OO designs are also f lexible and
extensible. But, most of the world’s data are stored in relational databases due to their
rigor and t he elegance of Structured Query Language (SQL).
A major development of the late 20th century is the development and adoption of spatial
databases based on object-relational database technologies. These technologies blend the rigor of
relational databases along with the f lexibility of OO. Non-spatial data can be stored as before,
while the spatial features can be stored as objects also within the same table as the non-spatial
data. This removed barriers between an organization’s GIS and non-GIS databases and allowed
development of multi-tier spatial database architecture where applications such as GIS, spatial
analysis and transportation models all share the same database (albeit with specialized support
for spatial data storage, access and processing).
Despite early clarion calls for spatiotemporal databases in the GIScience literature (Langran
1992), these methods and technologies were slow in developing, mostly due to lingering but
now lifting constraints on data capture, storage and processing. Another dramatic development
for GIS-T in the 21st century is the development of spatiotemporal and moving objects data-
bases (methods for representing, storing and accessing data that changes its geometry frequently,
especially its geographic location). Due to their currency, we will return to these topics in the
section below on 21st century GIS-T.
2.2. ROUTES AND FLOWS WITHIN SPATIAL NETWORKS
In addition to spatial networks, fundamental to GIS-T are the things moving within spatial net-
works – people, freight and vehicles. In 20th century GIS-T, there was little individual-level
data: most mobility data were aggregate f lows and counts. Therefore, a core problem in
transportation science during that era was predicting movements within spatial networks. This
requires finding minimum cost network paths, with cost defined broadly (in terms of distance,
time, money and/or mental effort). This also includes routing a single moving object within a
spatial network or determining the joint f low of multiple entities subject to their mutual
interactions (i.e., congestion). As we will discuss below, a big change in the 21st century is
the ability to collect individual-level mobility data. Nevertheless, being able to solve for routes
and f lows within spatial networks still has practical and theoretical importance: it is a core prob-
lem in transportation science and engineering.
Minimum cost routes and f lows through spatial networks are core topics for GIS-T for
reasons at the intersection of computation and geography. There are many ways to process net-
work data; some are designed for general networks not sparsely connected spatial networks.
While many transportation students learn min-cost path methods such as Dijkstra’s algorithm,
this is only part of the story. Equally important is the network data processing methods that feed
these algorithms; this includes how to store, search and process data. Miller and Shaw (2001)
review fifteen different min-cost path algorithm implementations, including eight different
versions of the Dijkstra algorithm; these differ based on how to select and manage the set of can-
didate nodes for the minimum cost path. Some perform well for spatial networks, and others
182 GIS-T in the 21st Century
© 2015 The Author(s)
Geography Compass © 2015 John Wiley & Sons Ltd
Geography Compass 9/4 (2015): 180–189, 10.1111/gec3.12204
perform better for some spatial networks but not others. However, it is typically unclear what
shortest path methods are being used in proprietary GIS software.
Understanding the nature of shortest path algorithms also leads to the essential insight that
computers are not magical: there are limitations on how fast they can process and what they
can solve. Two key transportation routing problems, the traveling salesman problem (TSP) and
the vehicle routing problem (VRP), are exemplars: they are simple to understand but cannot be
solved exactly (optimally) for even modest problems within a reasonable amount of time.
The TSP involves finding the shortest tour through a set of locations that returns to the start;
VRP is the multi-vehicle version of the TSP. Since these are very practical problems, we use
heuristics or shortcuts that generate a suboptimal but good solution quickly. These problems
occur in geographic space or spatial networks, therefore these heuristics often involve spatial rea-
soning (such as adding the nearest neighbor of the current tour or forming spatial clusters of stops
before routing). Similarly, there are spatial heuristics for speeding up the performance of shortest
path algorithms by limiting the size of the network to be processed based on the locations of
nodes. Spatial networks (and spatial processes more generally) are well-suited for parallel com-
puting and cyberinfrastructure since transportation systems are distributed and local in nature
and therefore can be decomposed by proximity in space and direction/nearness in time.
In Miller and Shaw (2001), we spent a considerable amount of time on solving for network
f low equilibrium patterns using optimization techniques. This includes the user optimal pattern
(everyone is on their minimum cost path, but overall cost is not minimized) and system optimal
pattern (the overall cost is minimized, but not everyone is on their minimum cost path). This is
core knowledge to GIS-T and indeed all of transportation science: the nature of these equilibria
has deep implications for theory and policy (e.g., how do we achieve a SO if some people can-
not use their min-cost path – coercion or persuasion?). These methods are also useful for plan-
ning over medium to long-term time horizons, particularly for infrastructure. They also serve as
the theoretical foundation of equilibrium travel demand and land-use/transportation models:
topics of considerable interest in the 20th century. As we will discuss below, a shifting policy
context for transportation combined with newly available data and computational power has
changed the focus of GIS-T.
2.3. SPATIAL ANALYSIS AND MODELING
Much of the power of GIS is at the front-end and back-end of any project. At the front-end,
GIS allows data integration and management based on geographic location. This sounds simple,
but it was very difficult prior to GIS. At the back-end, GIS facilitates mapping of data and results
for exploration, analysis and communication purposes. A map is an ancient and profound
communication technology with capabilities that have recently expanded to include anima-
tions, interactivity and exploration – the integration of cartography and scientific visualization
referred to as geovisualization (MacEachren et al. 2004). These basic spatial data management
and mapping capabilities alone would justify the use of GIS in transportation.
GIS also brings a set of spatial analytic capabilities that are useful for transportation. Basic spa-
tial analysis functionality includes the ability to query and manipulate georeferenced data based
on geometric, topological and set-based predicates, such as selecting all parcels within 500 feet of
a proposed transportation corridor. It also includes capabilities for processing fields such as
terrain to generate products such as hydrological networks, transportation corridors and
viewsheds. Core spatial analytical functions in GIS also include basic network analysis capabili-
ties such as minimum cost paths and routing. GIS also enables spatial analysis and modeling
through its capabilities to handle georeferenced spatial data and visualize results. Other spatial
analytic issues that are relevant to transportation include dealing with modifiable areal units such
GIS-T in the 21st Century 183
© 2015 The Author(s)
Geography Compass © 2015 John Wiley & Sons Ltd
Geography Compass 9/4 (2015): 180–189, 10.1111/gec3.12204
as census geography and traffic analysis zones, f low sampling at cordons and within networks,
capturing spatial dependence and spatial heterogeneity in travel and location models and alter-
native projections of geospace to represent travel environments and outcomes for visualization
and analysis (Miller 1999).
Fueled by GIS and increasingly available georeferenced data, the late 20th century witnessed
the rise of disaggregate spatial statistics such as local indicators of spatial association and
geographically weighted regression (Anselin 1995; Brunsdon, Fotheringham and Charlton,
1998). These capture both spatial dependency and heterogeneity, allowing relationships to be
disaggregated to the level of the spatial reporting unit. Transportation applications of these tech-
niques grew in the 21st century with wider recognition of their use for analyzing a wide range of
transportation and urban phenomena (Páez and Scott 2005).
2.4. 20TH CENTURY TRANSPORTATION APPLICATIONS
GIS-T in the 20th century was mostly about aggregate f lows within networks between large
geographic units such as traffic analysis zones (TAZs). For example, the discussion of transpor-
tation planning in Miller and Shaw (2001) included topics such as TAZ design, the four-step
approach to travel demand modeling (and its weaknesses), equilibrium transportation models
and artificial neural networks for predicting f lows. Similarly, discussion of hazards and environ-
ment focus on translating f lows into environmental impacts; measuring risk for network arcs;
analyzing accidents as points, lines (network segments) or polygons; locating and routing
emergency facilities within networks; and aggregate-level evacuation analysis. GIS-T to support
logistics focused on georeferenced data integration, network optimization and decision support.
By the turn of the 21st century, the dominance of aggregate transportation models was
challenged not only by new capabilities for data collection and individual-level simulation. Also
challenging this dominance is a changing empirical and policy context for transportation.
3. GIS-T in the 21st Century
3.1. THE CHANGING CONTEXT FOR TRANSPORTATION
Major global trends are shaping 21st century transportation. One trend is the increasing global-
ization of economies and societies that pushes for faster and more efficient transportation sys-
tems to move people and goods around the globe. Another global trend is the continuing
growth, urbanization and motorization of world population. Population growth is stalling in
the Global North (North America, Europe and Japan) but continuing in the Global South
(especially sub-Saharan Africa and Southeast Asia). The Global North faces challenges associated
with the changing needs of an aging population, while the Global South faces challenges of
more people and especially more automobiles as societies grow more aff luent. Almost every-
where in the world, people are crowding into cities: for the first time in human history, more
people live in cities than rural areas, a trend that is accelerating (Sperling and Gordon 2009).
Technological changes are also shaping transportation. Evolution from animal-powered
transportation to rail transport, automobile transport and air transport significantly changed
human mobility and interaction patterns. Although we do not anticipate a completely new
mobility technology in the foreseeable future, other technological advancements have caused
major changes to human activities. The most notable changes are due to information and commu-
nication technologies (ICTs) such as the Internet and mobile phone and location-aware technologies
(LATs) such as the global positioning system, mobile phones and Wi-Fi positioning techniques.
ICTs have changed the ways our economic systems operate and how people carry out their
184 GIS-T in the 21st Century
© 2015 The Author(s)
Geography Compass © 2015 John Wiley & Sons Ltd
Geography Compass 9/4 (2015): 180–189, 10.1111/gec3.12204
activities and interact with other people (Shaw 2014). LATs and ICTs are also not just for
humans: the rise of vehicle-to-vehicle (V2V) communication and the Internet of Things suggest
future transportation systems where many mundane activities (including driving) occur
automatically.
The policy context for transportation has also changed in the 21st century. For most of the
20th century, transportation policy was dominated by a “predict and provide” paradigm
(Bannister 2007). Planners sought normative (optimal) solutions based on aggregate models that
assume transportation systems achieved steady-state equilibrium and typically dictated large-scale
interventions. This suited the needs of the 20th century ( growth and expanding infrastructure) as
well as its mindset (cheap energy and an accommodating environmen t). These grand plans
neglect the small spaces and local movements that comprise everyday life, with the
consequence of inhumane urban settings (Batty 2012). While aggregate methods are still
well-suited for evaluating hard policies associated with infrastructure construction, management
and renewal, the early 21st century has seen an expansion in transportation policy to include
soft policies associated with travel demand management, such as information, persuasion
and incentives (Bamber g et al. 2011). These new policy demands require
individualist ic and dynamic approaches to analyzi ng and understandin g huma n mobility
(Miller 2015b; Shaw 2010).
3.2. MOBILITY AND ACTIVITY DATA
Due to advancements in ICTs and LATs, it is now feasible to collect large amounts of data from
a wide range of mobile sensors in real-time or near-real-time at high spatial and temporal
granularity. These datasets have potentials of providing us with unprecedented opportunities
to gain insight of mobility patterns and human behaviors that could be useful in transportation
planning, operation, maintenance and policy-making (Shaw 2011). It is not necessarily the
volume of these data that is revolutionary, but rather their ubiquity: they are the byproduct
of daily life, allowing naturalistic observation of mundane activities and serendipitous observa-
tion of unusual events (Miller and Gooldchild, 2014).
An increasing amount of human activities now take place in virtual space via ICT (Shaw and
Yu, 2009). People now can enjoy freedom of ordering items at amazon.com almost anytime
and anywhere via a smart mobile device and Internet connection and wait for the items shipped
to our home by a delivery company. Online shopping has inf luenced not only how individuals
carry out their shopping behaviors but also the logistics systems that deliver products to con-
sumers. Many companies now collect detailed data of customers’ shopping behaviors either at
retail stores or online. It also is common for companies to keep track of how and when items
are shipped throughout supply chain systems.
Social interaction patterns are changing due to widespread online social networking services
such as Facebook, Twitter, WhatsApp and Snapchat. To what extent these online social net-
work activities are inf luencing our physical travel patterns remains an unanswered research
question. How can we use various kinds of tracking data in a GIS environment to help us better
understand mutual interactions between our activities in virtual space and physical space that
would have important implications to transportation needs? This raises some fundamental
GIS research questions such as “how should we represent locations in a virtual space?” and
“how do we handle interactions between virtual space and physical space?” GIS-T in the
21st century therefore must tackle research challenges of dealing with human activities in virtual
space and their relationships with the activities in physical space if we want to use GIS-T to help
us design future transportation systems that can better serve the changing needs of our society
(Shaw 2010).
GIS-T in the 21st Century 185
© 2015 The Author(s)
Geography Compass © 2015 John Wiley & Sons Ltd
Geography Compass 9/4 (2015): 180–189, 10.1111/gec3.12204
3.3. GIS SOFTWARE FOR SPATIAL ANALYSIS AND MODELING IN THE 21ST CENTURY
As indicated in Miller and Shaw (2001), analysis plays a critical role in transportation studies.
Network analysis, travel demand analysis, facility location analysis and supply chain optimization
are among transportation analysis topics that have been widely pursued. GIS, on the other hand,
have developed a number of spatial analysis tools and incorporated some transportation analysis
functions into GIS software packages to support transportation and other application needs.
GIS technology still often falls short of offering an integrated environment that can effectively
and efficiently manage, analyze and visualize the diverse spatiotemporal data sources such as
dynamic traffic f low counts, tra cking data of individual movi ng objects (e.g., vehicles, pe r-
sons and parcels), video streams from sensors, individual activity data (shopping data, online
social media data, mobile phone tracking data, etc.) with the range of data coll ected by the
government agencies (t ransportation inventory data, real-time monitoring data, land-use
data, census data, etc.). The strategy of representing differ ent datasets as s tatic snapshot
GIS map layers is no longer sufficient to deal with the increasing amount of real-time and
dynamic data. One of the major challenges for GIS-T in the 21st century therefore is to
effectively and efficie ntly handle diverse spatiotemporal and dynamic datasets in an inte-
grated environment. This spatiotemporal GIS-T environment may need to support different
data models for various datasets; however, it should allow users to cr oss-reference different
datasets seamlessl y (Yin and Shaw, 2015in ).
Proprietary GIS software is developed for representation, management, analysis and visuali-
zation of geographic data in support of a wide range of applications. These GIS packages
normally are not designed to offer functions to support all transportation planning, operation,
maintenance and evaluation needs. Development of GIS-T therefore requires interactions
between researchers and practitioners in both transportation and GIS to identify how transpor-
tation can benefit from GIS and how GIS can best serve transportation. As technology advances
and transportation needs change over time, GIS-T will continue to evolve at a relatively fast rate
that ref lects the dynamic nature of today’sworld.
Open GIS is an emerging paradigm for GIScience that encompasses open data, software,
hardware, standards, collaboration, publication, funding and open education/learning. Open
GIS offers technology-led opportunities for addressing the spatiotemporal data deluge,
application-led opportunities for confronting a complex and rapidly changing planet, greater
capabilities for citizen science and crowdsourcing to solve mobility and accessibility
problems, and new approaches to educating the transportation professional. Challenges to
Open GIS include an academic reward system that favors closed science, intellectual prop-
erty issues, the potential disruption to power relationships and questions regarding induced
resource consumption and sustainability (Sui 2014). Open GIS can be more nimble and
responsive than proprietary GIS, offering better support for emerging transportation applica-
tions in the 21st century.
3.4. NEW TRANSPORTATION APPLICATIONS
Not surprisingly, we are seeing a new emphasis on the individual in 21st century transportation
applications of GIS-T. This includes the renaissance of activity-based analysis (ABA) approaches to
travel demand and transportation modeling and new applications to support customized
mobility services.
ABA treats transportation as a means to facilitate individual-level activities in space with
respect to time. ABA addresses the fundamental behaviors that determine travel and therefore
is better for supporting soft policies such as ridesharing, congestion pricing and complex trip
chaining. ABA also integrates better with methods for estimating environmental impacts of
186 GIS-T in the 21st Century
© 2015 The Author(s)
Geography Compass © 2015 John Wiley & Sons Ltd
Geography Compass 9/4 (2015): 180–189, 10.1111/gec3.12204
transportation such as individual-level emission models. It is also better at capturing the complex
interactions between mobility and information/communication technologies (Miller 2014).
Finally, since activity-based approaches are individualistic, they recognize social diversity and
are better suited for examining the differential impacts of travel demand management strategies
(Algers, Eliasson and Mattsson 2005; Malayath and Verma 2013; Miller 2015b). Major types of
activity-based models include econometric models, computational process models,
microsimulation methods and agent-based models (Buliung and Kanaroglou 2007; Miller
2014; Timmermans et al., 2002).
ABA theories and methods in transportation were well-established in the 20th century with
roots in time-use studies and time geography (Ellegård and Svedin 2012). However, their
computation-hungry and data-hungry nature could not be satisfied until the 21st century. In
addition, the increasing speed and complexity of human activities and the changing policy con-
text for transportation favors disaggregate over aggregate approaches. GIS is central to ABA due
to the essential nature of georeferenced data: indeed, GIS is even more essential to ABA than
aggregate approaches such as the four-step model due to greater needs for high-fidelity
representation of transportation infrastructure and services, as well as socio-economic data and
environmental data.
With the high-speed data communication for mobile devices, means of providing transpor-
tation services are going through a transformation stage. Most transportation services in the past
were designed as fixed-route and fixed-schedule services because there was no feasible means to
stay in touch with transportation users and keep them informed. The development of demand
responsive transit services in the past operated mainly as an advanced reservation system rather
than a (near) real-time demand responsive system. This has changed with smartphone apps
and on-demand mobility services such as Uber and Lyft. Many bikeshare systems also have
real-time station status apps, and public transit services are sharing their schedules and vehicle
GPS feeds with developers and the public. These online apps provide better means to match
supply and demand in (near) real time and could contribute to a more sustainable transport
system by allowing users to stitch together appropriate mobility services and depend less on
private automobile ownership. Other online apps allow users to check current highway traffic
conditions; the number of parking spaces available in a neighborhood; and navigation assistance
to vehicles, pedestrians and services to people with mobility, vision or hearing disabilities.
The convergence of location-aware technologies, sensors and machine intelligence is leading
to driverless vehicles, including automobiles for humans (e.g., Google) and unmanned aerial
vehicles for package delivery (e.g., Amazon, UPS) that could introduce fundamental changes
to how transportation systems operate. For example, self-driving vehicles can make
automobile-based mobility safer and more accessible to the elderly and people with different
abilities. It can also allow commutes to be productive and entertaining rather than robotic
and boring. Since a private automobile is vacant and stationary for much of its existence,
combining self-driving cars with car-sharing can allow current levels of mobility with fewer
private automobiles (Thrun 2013). However, more freely availability personal mobility may
also induce further sprawl and leap-frog development, with consequent negative impacts on
sustainability due to increased energy consumption associated with large, detached homes and
dispersed infrastructure. The appropriate integration of customized mobility services with col-
lective public transit is a key transportation scientific and policy question for the 21st century.
4. Conclusion
Transportation analysis and models have traditionally focused on data aggregated by geographic
units and over a relatively long time interval (e.g., spatial interaction models to estimate O–D
GIS-T in the 21st Century 187
© 2015 The Author(s)
Geography Compass © 2015 John Wiley & Sons Ltd
Geography Compass 9/4 (2015): 180–189, 10.1111/gec3.12204
f lows between traffic analyze zones in a day) due to constraints on data collection and compu-
tational power. Fostered by increasing capabilities for collecting, storing, processing and
analyzing individual-level mobility data, as well as changing empirical and policy context for
transportation, the 21st century is witnessing a major shift in GIS-T away from aggregate f lows
and static analysis to individual activities and spatiotemporal analysis. Although there are
enduring core principles, and some conventional transportation analysis methods and models
will remain useful to address some transportation needs, a major challenge that GIS-T faces is
how to analyze the big and dynamic datasets such that we can gain additional insight of travel
behaviors and transportation needs.
While the scientific, policy and planning challenges for 21st century GIS-T are formidable,
equally challenging and vital are educational issues associated with the new mobility science.
Many transportation education curricula today are still heavily based on the methods and
models developed during the mid-20th century. We do not have an appropriate curriculum
in either transportation or GIS to educate transportation students and practitioners on how to
analyze and model big dynamic datasets. This is an issue that needs immediate attention in
the GIS-T community.
Short Biographies
Harvey J. Miller is the Bob and Mary Reusche Chair in Geographic Information Science in the
Department of Geography at The Ohio State University in Columbus, OH, USA. He also has
affiliations with the Center for Urban and Regional Analysis, the Institute for Population Re-
search, and the Department of City and Regional Planning. His research and teaching interests
fall in the intersection between geographic information science and transportation science. He is
particularly concerned with how people use mobility and communications technologies to al-
locate scarce time among activities in geographic space – a perspective known as time geogra-
phy. He is also interested in the social dimensions of transportation and the implications of
human mobility and accessibility for sustainable transportation, livable communities, and public
health.
Shih-Lung Shaw is an Alvin and Sally Beaman Professor and an Arts and Sciences Excellence
Professor in the Department of Geography at the University of Tennessee, in Knoxville, TN,
USA. His research and teaching interests include transportation geography, geographic infor-
mation science (GIScience), and spatio-temporal analysis, especially with topics related to trans-
portation planning and modeling, air transportation, time geography, effects of information and
communications technologies (ICT) on human activity and travel patterns, spatiotemporal anal-
ysis of human dynamics, space-time GIS, and GIS for transportation (GIS-T).
Note
*
Correspondence address: Harvey J. Miller, Department of Geography, The Ohio State University, 1036 Derby Hall/154
N. Oval Ma, Columbus, OH 43210, USA. E-mail: miller.81@osu.edu
References
Algers, S., Eliasson, J. and Mattsson, L.-G. (2005). Is it time to use activity-based urban transport models? A discussion of
planning needs and modelling possibilities. Annals of Regional Science 39, pp. 767–789.
Anselin, L. (1995). Local indicators of spatial association—LISA. Geographical Analysis 27, pp. 93–115.
Bamberg, S., Fuiji, S., Friman, M. and Gärling, T. (2011). Behaviour theory and soft transport policy measures. Transport
Policy 18, pp. 228–235.
Bannister, D. (2007). The sustainable mobility paradigm. Transport Policy 15, pp. 73–80.
Batty, M. (2012). Smart cities, big data. Environment and Planning B 39, pp. 191–193.
188 GIS-T in the 21st Century
© 2015 The Author(s)
Geography Compass © 2015 John Wiley & Sons Ltd
Geography Compass 9/4 (2015): 180–189, 10.1111/gec3.12204
Brunsdon, C., Fotheringham, S. and Charlton, M. (1998). Geographically weighted regression. Journal of the Royal Statistical
Society: Series D 47, pp. 431–443.
Buliung, R. N. and Kanaroglou, P. S. (2007). Activity–travel behaviour research: conceptual issues, state of the art, and
emerging perspectives on behavioural analysis and simulation modeling. Transport Reviews 27 (2), pp. 151–187.
Ellegård, K. and Svedin, U. (2012). Torsten Hägerstrand’s time-geography as the cradle of the activity approach in transport
geography. Journal of Transport Geography 23, pp. 17–25.
Goodchild, M. F. (1992). Geographical information science. International Journal of Geographical Information Systems 6,
pp. 31–45.
Langran, G. (1992). Time in GIS. London: Taylor and Francis.
MacEachren, A. M., Gahegan, M., Pike, W., Brewer, I., Cai, G., Lengerich, E. and Hardistry, F. (2004) Geovisualization for
knowledge construction and decision support. Computer Graphics and Applications, IEEE,24(1),pp.13–17.
Malayath, M. and Verma, A. (2013). Activity based travel demand models as a tool for evaluating sustainable transportation
policies. Research in Transportation Economics 38, pp. 45–66.
Miller, H. J. (1999). Potential contributions of spatial analysis to Geographic Information Systems for Transportation
(GIS-T). Geographical Analysis 31, pp. 373–399.
Miller, H. J. (2014). Activity-based analysis. In: Fischer, M. and Nijkamp, P. (eds) Handbook of regional science. Berlin: Springer,
pp.705–724.
Miller, H. J. (2015b) Theories and models in transportation planning. In: Hanson, S., and Giuliano, G. (eds.) The geography of
urban transportation. 4th ed., in press.
Miller, H. J. and Goodchild, M. F. (2014). Data-driven geography. GeoJournal. DOI: 10.1007/s10708-014-9602-6
Miller, H. J. and Shaw, S.-L. (2001). Geographic Information Systems for Transportation: principles and applications. New York:
Oxford University Press.
Páez, A. and Scott, D. M. (2005). Spatial statistics for urban analysis: a review of techniques with examples. GeoJournal 61,
pp. 53–67.
Shaw, S.-L. (2010). Geographic information systems for transportation: from a static past to a dynamic future. Annals of GIS
16 (3), pp. 129–140.
Shaw, S.-L. (2011). Guest editorial – geographic information systems for transportation: an introduction. JournalofTransport
Geography 19 (3), pp. 377–378.
Shaw, S.-L. (2014) Geographic information systems in transportation. In: Garrett, M. (ed) Encyclopedia of transportation: social
science and policy. Thousand Oaks, CA: SAGE Publications. DOI: 10.4135/9781483346526
Shaw, S.-L. and Yu, H. (2009). A GIS-based time-geographic approach of studying individual activities and interactions in a
hybrid physical-virtual space. Journal of Transport Geography 17 (2), pp. 141–149.
Sperling, D. and Gordon, D. (2009). Two billion cars: driving towards sustainability. New York: Oxford University Press.
Sui, D. (2014). Opportunities and impediments for open GIS. Transactions in GIS 18, pp. 1
–24.
Thrun, S. (2013) Google’s original X-man: a conversation with Sebastian Thrun. Foreign Affairs. http://www.foreignaffairs.
com/discussions/interviews/googles-original-x-man (last accessed 9 March 2015)
Timmermans, H. J. P., Arenze, T. and Joh, C.-H. (2002). Analyzing space-time behavior: new approaches to old problems.
Progress in Human Geography 26 (2), pp. 175–190.
Winter, S., Sester, M., Wolfson, O. and Geers, G. (2010). Towards a computational transportation science. SIGMOD Record
39, pp. 27–32.
Yin, L. and Shaw, S.-L. (2015). Exploring space-time paths in physical and social closeness spaces: a space-time GIS approach.
International Journal of Geographical Information Science. DOI: 10.1080/13658816.2014.978869
GIS-T in the 21st Century 189
© 2015 The Author(s)
Geography Compass © 2015 John Wiley & Sons Ltd
Geography Compass 9/4 (2015): 180–189, 10.1111/gec3.12204
