Seediscussions,stats,andauthorprofilesforthispublicationat:

https://www.researchgate.net/publication/263078531

Space-TimeRepresentationandAnalytics

Article

in

AnnalsofGIS·January2014

DOI:10.1080/19475683.2013.862301

CITATIONS

3

READS

88

3authors:

MayYuan

UniversityofTexasatDallas

67

PUBLICATIONS

1,019

CITATIONS

SEEPROFILE

AtsushiNara

UniversityofOklahoma

16

PUBLICATIONS

94

CITATIONS

SEEPROFILE

JamesBothwell

UniversityofOklahoma

7

PUBLICATIONS

15

CITATIONS

SEEPROFILE

Allin-textreferences

underlinedinblue

arelinkedtopublicationsonResearchGate,

lettingyouaccessandreadthemimmediately.

Availablefrom:MayYuan

Retrievedon:14April2016

This article was downloaded by: [University of Oklahoma Libraries], [Dr May Yuan]
On: 16 April 2014, At: 11:49
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Annals of GIS

Publication details, including instructions for authors and subscription information:

http://www.tandfonline.com/loi/tagi20

Space–time representation and analytics

May Yuan

a

, Atsushi Nara

a

& James Bothwell

a

a

Center for Spatial Analysis, University of Oklahoma, 301 David L. Boren Blvd, Norman, OK

73019, USA
Published online: 10 Jan 2014.

To cite this article: May Yuan, Atsushi Nara & James Bothwell (2014) Space–time representation and analytics, Annals of GIS,
20:1, 1-9, DOI:

10.1080/19475683.2013.862301

To link to this article:

http://dx.doi.org/10.1080/19475683.2013.862301

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at

http://

www.tandfonline.com/page/terms-and-conditions

Space

–time representation and analytics

May Yuan

*

, Atsushi Nara and James Bothwell

Center for Spatial Analysis, University of Oklahoma, 301 David L. Boren Blvd, Norman, OK 73019, USA

(Received 1 October 2013; accepted 30 October 2013)

This paper discusses philosophical and methodological considerations of space

–time representation and analytics. Central to

our premise is that how we conceptualize space and time has profound in

fluence on the way in which we represent and

analyse spatiotemporal data. Four approaches in space

–time representation are discussed based on two sets of space–time

dichotomies: absolutism versus relationalism and realism versus idealism. We elaborate on the philosophical inquiries
related to the dichotomies and discuss approaches taken in GIScience research. Since most GIS studies to date align well
with the absolutism

–realism approach to space–time representation, we discuss examples that adopt absolutism–idealism,

relationalism

–realism, and relationalism–idealism approaches. We argue that multiple perspectives of space and time can

bring rich insights into understanding geographic phenomena. All four approaches of space

–time representation and

ensuing analytical methods should be broadly explored in GIScience research.

Keywords: space

–time; representation; analytics

1.

Introduction

Space and time, hereafter space-time, are fundamental yet
complex subjects that have stimulated fascinating debates
among philosophers, physicists, geographers, historians
and scholars in many other

fields. Most of the ontological

and epistemological questions concerning space and time
from Kant and his fellow philosophers (Janiak

2012

)

1

remain

relevant

to

geographic

information

science

(GIScience) research today:

(1) Question about the ontology and substance

–prop-

erty metaphysical framework: Is space

–time a sub-

stance in its own right or a property of some
substance? The question inquires the relationship
of space

–time to physical objects, contrasting

views of absolutism (viz. space

–time points

exist)

and

relationalism.

Absolutism

views

space

–time as an object-independent framework,

whereas relationalism considers space

–time as the

order of possible relations among objects.

(2) Question about the origin of space

–time represen-

tation: Where do our ideas about space and time
come from? Do the ideas of space and time origi-
nate from the perceptions and experiences of spa-
tial or temporal relations? Is there any distinct
issue about space

–time from ordinary physical

objects? While we can represent space as empty
of objects, can we represent the absence of space

or time? Are space and time causally inert and
therefore imperceptible? If so, it would be impos-
sible to represent space and time. Is the origin of
space

–time representation constrained by spatio-

temporal objects? Or is it possible to have a repre-
sentation of space and time themselves? Are space
and time dependent upon the mind for their exis-
tence? Realists consider relations are mind-inde-
pendent;

for

idealists,

relations

are

mind-

dependent. Is the dependence suggested by the
origin

or

the

content

of

our

space

–time

representation?

(3) Question about the content of space

–time repre-

sentation: Does our idea about space re

flect what

we know from Euclidean Geometry? Can space

–

time be an empirical concept derived from outer
experiences? Can the content of a representation
guide us to its possible origin? Can the origin of
a representation lead to its potential content? If
we represent space containing an in

finite number

of places within it, each of the places would also
have an in

finite number of subordinate concepts

about space. Hence, a concept contains an in

finite

number of concepts, each of which has an in

finite

number of constituents of sub-concepts and
recursive sub-concepts, and eventually the con-
cept of space lacks determinant content and
therefore is deemed incomprehensible. If our

*Corresponding author. Email:

[e-mail]

Annals of GIS, 2014
Vol. 20, No. 1, 1

–9, http://dx.doi.org/10.1080/19475683.2013.862301

© 2013 Taylor & Francis

Downloaded by [University of Oklahoma Libraries], [Dr May Yuan] at 11:49 16 April 2014

representation of space is not a concept, is it an a
priori intuition?

These philosophical inquiries nurture our thinking about
space

–time, and our space–time conceptualization influ-

ences how we construct geographic representation and
consequently constrains what analytics we can perform
and what insights we can gain from research. For philo-
sophers, a representation is

‘an image, concept, or thought

in the mind as representing an object or state of affairs in
the world

’ (c.f. Oxford English Dictionary Third Edition,

December 2009). The object of a representation captures
semantic properties, such as content, truth-conditions and
truth-value, philosophically speaking. In the context of
GIScience, space

–time representation encapsulates in

forms of data objects the essential characteristics of enti-
ties, facts, concepts, and states of affairs on Earth (or
simply things on Earth).

Things on Earth, abstract or concrete, can be challen-

ging to conceptualize and consequently dif

ficult to repre-

sent. On the other hand, scienti

fic research relies upon

observations and therefore is bound by means that we can
express our observations, quantitatively or qualitatively.
Taking the absolutist

’s view, GIScientists represent space

as a collection of lattice points, for example, with semantic
properties at locations and then expand the lattice orthogon-
ally to a cube to incorporate the temporal dimension. Such a
space

–time cube representation can be traced back to

Hägerstrand

’s time geography (Hägerstrand

1970

) and is

shown effective in information visualization (Kristensson
et al.

2009

), spatiotemporal analysis and synthesis (Nakaya

2013

), visual exhibition exploration (Windhager and Mayr

2012

), space

–time data modelling and mathematical mod-

elling (Gatalsky, Andrienko, and Andrienko

2004

). In fact,

most space

–time research in GIScience and related fields

adopts the view of absolutism that space

–time is an object-

independent framework (e.g. a cube), existing on its own as
a container populated with space

–time points or objects.

The absolutist

’s view is often accompanied with the

idea of realism in which things on Earth are perceived and
experienced as

fields and objects. Consequently, geo-

ontology studies decipher categories of geographic phe-
nomena to various kinds of

fields and various kinds of

objects in space and time (Galton

2003

) or theorize the

ontological hierarchy of geographic information that con-
nects existence, observable, class, simple object, compo-
site object, function, and purpose levels of representations
(Couclelis

2010

). Human experiences position objects and

fields in the pre-defined space–time framework. The
values of space and time for a given thing (object or
field) depend on the location and chronological time of
observations. In other words, space and time are pre-
de

fined, regardless of the existence of objects and fields.

An alternative

‘bottom-up thinking’ of geographic repre-

sentation starts with the idea of geo-atom, the very primitive

form of geographic information (Goodchild et al.

2007

). A

geo-atom is a space

–time point with semantic properties and

associated values for the properties. When geo-atoms are
bounded by the same identity over space and time, these
geo-atoms represent a geo-object. When geo-atoms are
bounded by the same property over space and time, these
geo-atoms represent a geo-

field. Various means of aggrega-

tion and agglomeration of geo-atoms in space and time form
representations of complex geographic phenomena. When
the ontology of geographic information is the focus, such a
bottom-up approach is realized by observation-driven ontol-
ogy engineering according to sensors, observation proce-
dures and observations (Janowicz

2012

).

Relationalism and idealism of space and time are yet to

flourish in GIScience research. Some research in GIS data
modelling touched on the idea that vector data models only
include space occupied by geometric objects and therefore
dismiss the possibility of empty space, hence that space can-
not be a substance in its own right (Couclelis

1999

). However,

the underlying geographic coordinate framework very much
grounds the space where the objects occupy, so, in essence,
vector data models remain fundamentally absolutistic and
substantivalistic. From a relational perspective, the existence
of space

–time depends on possible objects and relations. GIS

examples of space

–time relationalism are address-matching

and routing, in which locations and paths can only exist in
relation to street networks. Idealism views space

–time as an a

prior intuition, i.e. a non-empirical, singular, immediate repre-
sentation of space that cannot be derived from outer experi-
ence but arises from the mind-dependent perception of spatial
relations. While virtual space, social space and political space
may not be intuitive, these spaces are idealistic in nature.
Regarding time, psychological time and parallel time are
examples of idealistic time which are not elements of reality
but relations added by the mind. A growing number of critical
GIS studies take feminist and qualitative approaches that are
mostly aligned with the idealistic view of space and time
(Kwan

2007

; Cope and Elwood

2009

). With the emphases

of knowledge production different from the objective and
quantitative trajectories, critical GIS pushes for the values of
social theory and neogeography in support of local epistemol-
ogy on life-worlds (Sheppard

2005

).

This paper examines the four quadrants of space

–time

representation (

Table 1

) in GIScience: absolutism

–realism,

absolutism

–idealism, relationalism–realism and relational-

ism

–idealism, and how representation influences space–

time analytics. While there are rich research programmes in
critical GIS, this paper limits discussions from a scienti

fic

and computational perspective on space

–time representation.

Ontological issues from social theory and critical theory
perspectives on realism, empiricism and idealism are beyond
the scope of the discussions here. The premise is that our
thinking and questioning about space

–time inspire the way

in which we construct representation, and furthermore, con-
strain what analytics we can do and conclusions we can draw

2

M. Yuan et al.

Downloaded by [University of Oklahoma Libraries], [Dr May Yuan] at 11:49 16 April 2014

(

Figure 1

). Absolutism

–realism of space and time is well

represented in a wide range of GIS studies, such as in land-
use and land-cover change and spatial ecology. It may be fair
to say that the dominant majority of GIS projects assume
absolute space

–time based on a Cartesian coordinate system

and solar calendar with clock time. The other three views of
space

–time deserve additional attention. The following three

sections will present a study in each section to highlight
absolutism

–idealism, relationalism–realism and relational-

ism

–idealism, respectively, in space–time representation

and analytics. In this paper, the purpose of these examples
is to illustrate the conceptual differences in treatments of
space and time and how a novel conceptualization of space
and time will lead to effective representation and analytics
for new insights about geographic dynamics. Technical
details on algorithmic and computational procedures are
beyond the scope of the paper. Readers who are interested
in the technical details should consult the cited references in
the respective section. This paper will conclude with a synth-
esis of

findings from the three studies and new perspectives

to space

–time representation and analytics.

2.

Absolutism

–idealism of space–time

Absolutism

–idealism posits that space–time is an object-

independent framework for mind-dependent object relations.
The space

–time framework is conceptualized as a container

to position and reference things of interest. Environmental
remote-sensing data and model outputs are good examples,
in that data are recorded at a pre-de

fined regular grid space

and regular time intervals. Space

–time is conceptualized as a

series of time-stamped spatial grids and at each grid point, a

property value is recorded for the theme of interest, such as
temperature. Animation is a common technique to visualize
temporal changes captured in the data, and map algebra and
matrix calculations by mathematical and statistical opera-
tions at pre-de

fined space and time. An idealistic view will

look beyond space

–time data records to conceptualize

higher-level phenomenal constructs out of the recorded prop-
erty, such as heat waves out of temperature observations. By
conceptualizing and representing heat waves or heat

flows,

we enhance our intuited space

–time perception of the world.

Such abstract, mind-constructed objects can be very effective
in eliciting hidden patterns in space

–time data.

The use of heat

flows has been shown to be effective

in comparison of temperature projections between two
general circulation models: the CNRM-CM3 model by
Centre

National

de

Recherches

Meteorologiques

(CNRM), Meteo France and the CCSM3 model by
National Center for Atmospheric Research (NCAR) in
the United States (Bothwell and Yuan

2010

). Instead of

representing temperature values at grid cells and a given
time, the conceptualization of heat

flows results in flow

vectors across grid cells over time. Flow vectors may be
de

fined in different ways. One can be the direction and

shortest distance to the cell of the same value at the next
time-stamped snapshot. Another possible de

finition of

flow vectors could be that each grid cell records a vector
with direction to the adjacent cell with the largest tem-
perature gradient and the magnitude of the temperature
difference. Regardless of

flow vector definitions, the ima-

ginary vectors provide direct mapping to concepts like
expansion, retreat, divergence and convergence.

For global temperature projections, heat

flow vectors

moving northwards or southwards from the equator corre-
spond to warming effects (

Figure 2

). When they move

towards the equator, it suggests a cooling effect. When heat
flow vectors serve as the primitives in representing tempera-
ture over space and time, we can query

flow vectors based on

the moving direction of interest to reveal warming or cooling
dominant regions. For example,

‘heat flows’ moving towards

the poles (northwards in the northern hemisphere or south-
wards in the southern hemisphere) suggest a warming trend
(e.g. a and b in

Figure 2

). Accordingly,

Figure 2

shows that

the NCAR model suggests a much wider spread of warming
in the southern hemisphere during 2030

–2090 than what is

predicted by the CNRM model. We can also calculate spatial
variability of moving directions and temperature differences

Figure 1.

Conceptualization, representation and analytics.

Table 1.

Four quadrats of space

–time representation (Janiak

2012

).

1.

Absolutism

Relationalism

Realism Space is an object-independent framework for object relations.

Relations are mind-

independent

Space is the order of possible relations among objects

Relations are mind-

independent.

Idealism Space is an object-independent framework for object relations.

Relations are mind-dependent

Space is the order of possible relations among objects.

Relations are mind-dependent

Annals of GIS

3

Downloaded by [University of Oklahoma Libraries], [Dr May Yuan] at 11:49 16 April 2014

to identify mixing zones. Moreover, we can analyze the
direction and magnitude of changes in

flow vectors to reveal

areas of transition zones where temperature change goes from
slow to fast, convergent zones where temperature change
converges to a hot spot or cold spot, and divergent zones
where warm or cold temperature expands outwards. By relat-
ing coastlines to

flow vectors, we can also determine land–sea

temperature relationships, such as a cooling effect from sea to
land. Based on c and d in

Figure 2

, CNRM model suggests a

stronger regional moment of heat

flows easterly, and sea to

land

flows in east Brazil, as opposed to land to sea movement

suggested by NCAR model.

This example adopts an absolute space

–time framework

in which space is set at an absolute grid of 1° latitude × 1°
longitude cells, and time is set at 60 years of monthly
temperature estimates. Space

–time relations are constructed

idealistically based on the concept of heat

flowing from

high to low temperature across space over time. The object
‘heat flow’ is imaginary and constructed in mind as a
surrogate for understanding temperature change and its
patterns. Beyond mapping and visualization, each cell in
the grid is associated with the data that record the direction
and distance of movement. Therefore, query and analysis
can be performed to discern locations with fastest spreading
of warming or reveal pronounced differences between the
estimates from the two models.

3.

Relationalism

–realism of space–time

The relationalism

–realism view of space–time posits that

space

–time is the order of relations among objects, and

objects are observable and measurable, and therefore

mind-independent. Since space

–time depends on the existence

of the objects of interest, space

–time analysis can only work

where and when there are objects. Spatial and temporal refer-
ences of an object are relative to other known objects in the
system. Track analysis is used here to demonstrate the space

–

time representation from the perspective of relationalism and
realism. A conventional approach to track analysis puts a great
emphasis on trajectory density and clustering analysis (Kwan

2004

; Li, Han, and Kim

2006

; Palma et al.

2008

; Li et al.

2011

),

which follows the absolutism

–realism view. A relationalism–

realism approach will focuses on the relationship among tracks
to discern potential patterns or implications.

Using global positioning system (GPS) points taken

for an individual from 1 May to 31 October 2009, we
attempt to show a relationalism

–realism space–time

approach for track analysis. Detailed algorithms and ana-
lyses are available in Yuan and Nara (Forthcoming).

Figure 3

shows the 21,160 GPS points only during the

month of September for the individual. The massive GPS
points obscure meaningful movement patterns. Taking a
relationalism approach, we compare stops made on a day
to those closest stops made on the previous day and
analyse accumulative distance of the stop pairs for differ-
ent days of the week over the 6-month period (1 May
2009

–31 October 2009). If the individual has a weekly

routine of stops (such as going to work), there will be
some regularity in the

fluctuation of the accumulative

distance among closest stops the next day over time.

Figure 4

summarizes the result and shows the following.

(1) The individual stopped at nearby locations Saturday

to Wednesday and Thursday to Friday in general.

Figure 2.

Decipher temperature change over space and time using mind-dependent heat

flows. In (A) and (B), dark-shaded colour

corresponds to

flows to the south, and light-shaded colour, to the north. In (C) and (D), dark-shaded colour corresponds to flows to the

west, and light-shaded colour, to the east.

4

M. Yuan et al.

Downloaded by [University of Oklahoma Libraries], [Dr May Yuan] at 11:49 16 April 2014

(2) On Thursdays and Saturdays, the individual

tended to stop by locations away from stops
made on Wednesdays and Fridays.

Likewise, we can also compare track differences to discern
meaningful patterns. Because GPS points were taken at
different times, resampling was used to derive tracks of
locations at equal time intervals (

Figure 5

). We can com-

pare tracks taken weekly at the same day of the week to
examine, for example, if the individual generally took the
same journey every Monday. We can compare a track to
the one taken on the previous day to assess how the
individual might adhere to a daily commute pattern.

Figure 6

summarizes the track differences to the previous

day throughout the 6-month period. The darker shade in
the image corresponds to smaller differences in tracks. A
distinctive pattern is the repetitive pattern of two white
lines separated by a black line throughout the study per-
iod. These white-black-white patterns corresponded to
tracks taken on Thursdays, Fridays and Saturdays. That
the individual had days off on Thursdays and Fridays and
resumed work on Saturdays resulted in a large difference
in tracks between Wednesdays and Thursdays, similar
tracks on Thursdays and Fridays and again large differ-
ence between Fridays and Saturdays. In the early dates,
there was no clear pattern, which might be indicative of

Figure 3.

GPS points taken for an individual from 1 September 2009 to 30 September 2009 as an example.

Figure 4.

Individual stop comparison over a 6-month period (Adapted from Yuan and Nara, Forthcoming).

Annals of GIS

5

Downloaded by [University of Oklahoma Libraries], [Dr May Yuan] at 11:49 16 April 2014

the period in which the individual was new to the town
and was looking for work (the top circle in

Figure 6

). The

space

–time track image also suggests that the individual

took outings around 20:00 on workdays in four separate
weeks starting in late May and then in August, September
and October, but usually returned home by 22:00 (the four
small circles in

Figure 6

). Furthermore, the individual

went out frequently around midnight on Thursdays.

The track analysis above provides an example of the

relationalism

–realism approach, in which we considered

space

–time as the order of possible relations among GPS

observations to elicit meaningful patterns from the massive
data points. Speci

fically, we applied time as the first order

of relations among observations, and considered space
based on the distance among points and tracks to discern
further relations. Rather than investigating into the density
or clustering of GPS points or tracks, we mined the data to
reveal deviations and periodicities of stops and tracks over
time to understand movement patterns, detect irregular out-
ings and recognize possible emergence of new routines.

4.

Relationalism

–idealism of space–time

While still using space

–time as the order of possible

relations, a relationalism

–idealism approach examines

the relations among objects that are intuitive and mind-
dependent. As such, the space

–time representation relies

upon abstract constructs conceptualized in human minds
prior to the formation of space

–time concepts. Spatial

social network analysis is a good example of relational-
ism

–idealism approaches. A social network is a social

phenomenon with entities tied with interactions or inter-
dependence (Borgatti et al.

2009

; Carpenter, Li, and

Jiang

2012

). Space

–time in a social network is deter-

mined by the member activities in the network, and
space

–time makes the connections among the members

and

their

activities.

Social

network

data

analytics

assumes potential links among individuals and explores
the characteristics of the network as a whole or its
compositions (Aggarwal

2011

). Nevertheless, a social

network is an abstract construct that provides an intuitive
surrogate for analysis and reasoning of the social

Figure 5.

Resampling using a mean

filter to match locations along tracks for comparison.

Figure 6.

A temporal image of track differences (Adapted from Yuan and Nara, Forthcoming).

6

M. Yuan et al.

Downloaded by [University of Oklahoma Libraries], [Dr May Yuan] at 11:49 16 April 2014

phenomena of interactions or interdependence. Our
example below assumes that the existence of social net-
works of potential interactions may be identi

fied by

space

–time proximity of individuals.

We used GPS data from 2795 individuals in Oklahoma

from 23 February 2009 to 22 January 2011 to uncover
social networks of potential interactions (Yuan and Nara,
Forthcoming). Synchronous presence thresholds were set
at a space

–time proximity of 45 m and 30 minutes.

Relations among the individuals were determined by
their space

–time proximity. Only individuals with GPS

points recorded in the de

fined thresholds were considered

related and therefore included in the social network of
potential interactions. Among the 2795 individuals, GPS
points from 373 individuals fell within the space

–time

proximity that created 54 social networks of potential
interactions (

Figure 7

).

Figure 8

shows one of the largest networks that

involves 78 individuals. In the network, we can identify
three distinctive sub-networks with one-to-one (

Figure 8

A),

one-to-many (

Figure 8

B) and many-to-many (

Figure 8

C)

relations. In the one-to-one relation, ID1 and ID2 have a
strong relation, in which both trajectories and stops on days

of potential interactions were very similar. Such long
space

–time coexistence indicates kinship and strong friend-

ship. In the one-to-many example, a central node in the
network (ID3) had all weak relations with 21 individuals,
where each potential interaction event was in a short

Figure 8.

The estimated social network of potential interactions involving 78 individuals in Oklahoma and examples of their

trajectories, stops and potential interaction locations on days of their potential interactions.

Figure 7.

The estimated social networks of potential interac-

tions among 2795 individuals in Oklahoma based on their GPS
points. Each node corresponds to each individual and its shade
represents an individual

’s total potential interaction duration. The

thickness of the links corresponded to the frequency of synchro-
nized presence.

Annals of GIS

7

Downloaded by [University of Oklahoma Libraries], [Dr May Yuan] at 11:49 16 April 2014

duration. As shown in the

Figure 8

B, the ID3

’s trajectories

covered across a wide area; therefore, the individual had
more opportunities to interact with other individuals. This
weak relationship could be explained by a post person who
would have frequent interactions with people at many
delivery locations. In

Figure 8

C,

five individuals have

strong relations with each other, creating a pentagon-
shaped network. On the map, there is one potential inter-
action location pointed by an outlined grey arrow that is for
multi-individual congregation. This location could corre-
spond to their social activities, such as visiting a common
friend, hanging out at a bar, or attending meetings.

The social network example illustrates two key

notions of relationalism

–idealism with respective to

space

–time representation and analytics. First, the

assumed existence of social networks of potential interac-
tions sets the scope to which space and time should be
conceptualized and analysed. There are four different
kinds of interactions according to constraints of synchro-
nized vs. asynchronized space and time (Janelle

1995

).

The example only considers social networks of potential
interactions that are synchronized in space and time. If
asynchronized interactions are of interest, the treatment of
space and time in the analysis will need to be signi

ficantly

rethought. Therefore, the representation of space and time
depends upon the conceptualization of social networks of
interest in prior. Second, once the nature of social net-
works of interest is determined, space

–time is conceptua-

lized in the framework of object activities. Considerations
are only given to where and when recorded in GPS data.
Comparative measures of space and time determine the
relations among objects, and space

–time connections of all

related objects form social networks that can only be
idealized in human minds.

5.

Conclusions

Central to the paper is the idea that space

–time representation

sets the foundation for space

–time analytics and consequently

analytical outcomes. We revisited key philosophical questions
about space and time and explored GIS studies that were
aligned with different thoughts of space

–time representation.

Speci

fically, the three key questions were discussed on

the ontological basis, origin and content of space and time,
and these questions rami

fied themselves in two dichoto-

mies: (1) about the space

–time existence in terms of

absolutism and relationalism; and (2) about the origin
and content of space

–time in terms of realism and

idealism. Absolutism contrasts relationalism by the view
that space

–time exists in its own right (i.e. an object-

independent framework) rather than as properties of
physical objects (i.e. the order of possible relations
among objects). Realism contrasts idealism by the
view that space

–time and objects are mind-independent

(e.g. physical objects) rather than mind-dependent (e.g.

abstract

constructs).

The

combination

of

the

two

dichotomies resulted in four possible space

–time represen-

tations:

absolutism

–realism,

absolutism

–idealism,

relationalism

–realism and relationalism–idealism. GIS

research

commonly

adopts

the

absolutism

–realism

approach to space

–time representation. This is evident

that conventional GIS applications analyses, computes
and maps data in a Cartesian system with pre-de

fined

time intervals. Examples of such GIS applications include
thematic mapping, crime density analysis and site suitability
modelling. This paper offers three examples to demonstrate
the other three approaches of space

–time representation.

The absolutism

–idealism approach was demonstrated

by the representation of heat

flows as an integral concept

of space

–time–temperature change to compare predictions

from two general circulation models. Common approaches
would follow an absolutism

–realism framework to com-

pare temperature differences at the pre-de

fined locations

(e.g. grid cells or climate stations) over time. Using the
abstract concept of heat

flows that represent the magnitude

and direction of imagery movements along the maximum
temperature gradient at the pre-de

fined locations, we are

able to discern patterns that correspond to warming, cool-
ing and divergence or convergence of hot spots.

As an example of relationalism

–realism approaches,

we analyzed GPS points of an individual collected over
a 6-month period. Instead of density or clustering pat-
terns of the points and tracks, we related points and
tracks spatially and temporally to seek implications for
potential routine and incongruent activities. Space

–time

representation functioned as measures of similarity and
ordering of relations among the points and tracks.
Finally, we took a relationalism

–idealism approach to

elicit social networks of possible interactions based on
GPS data of 50 individuals. Relations among the indi-
viduals were determined by the proximity of their GPS
points. Space

–time set the proximity thresholds for pos-

sible interactions in building social networks. The
abstract concepts of social networks were conceived
prior to the conceptualization of space and time and
subsequently the analytical procedures to reveal the
potential links among the individuals.

Our premise posits that the four views of space

–time

representation

offer

profound

perspectives

for

GIS

research. Our conceptualization of space

–time leads us to

different strategies in space

–time representation and ana-

lytics, and consequently draws insights from multiple
perspectives. Currently, most GIS applications subscribe
to absolutism

–realism thinking of space and time and have

been productive in analysing objects or

fields from which

data were collected. What relationalism and idealism sug-
gest are rich suites of new possibilities by considering
abstract constructs of rich semantics that can be built
from space

–time data based on the relations among things

on Earth.

8

M. Yuan et al.

Downloaded by [University of Oklahoma Libraries], [Dr May Yuan] at 11:49 16 April 2014

Acknowledgements

Moreover, we can analyse the direction and magnitude of changes
in

flow vectors to reveal areas of transition zones where temperature

change goes from slow to fast, convergent zones where temperature
change converges to a hot spot or cold spot and divergent zones
where warm or cold temperature expands outwards.

Funding

This material is based upon work supported by the National
Science Foundation [grant number OCI 0941501] and National
Institute of Justice [award number 2010-DE-BX-K005]. Any
opinions,

findings and conclusions or recommendations

expressed in this article are those of the authors and do not
necessarily

re

flect the views of the National Science

Foundation or National Institute of Justice.

Note

1.

The questions are summarized from Stanford Encyclopedia
of Philosophy (SEP) on Kant

’s views on space and time

(

http://plato.stanford.edu/entries/kant-spacetime

/

accessed

July 30, 2013). SEP lists

five distinct questions. In this

paper, we merged the questions on ontology and the sub-
stance

–property framework as well as questions on the ori-

gin of space

–time representation and views of realism and

idealism for their highly related nature of inquiries.

References

Aggarwal, C. C. 2011.

“An Introduction to Social Network Data

Analytics.

” In Social Network Data Analytics, edited by

C. C. Aggarwal, 1

–15. New York: Springer.

Borgatti, S. P., A. Mehra, D. J. Brass, and G. Labianca. 2009.

“Network Analysis in the Social Sciences.” Science 323
(5916): 892

–895.

Bothwell, J., and M. Yuan. 2010.

“Apply Concepts of Fluid

Kinematics to Represent Continuous Space

–Time Fields in

Temporal GIS.

” Annals of GIS 16 (1): 27–41.

Carpenter, M. A., M. Li, and H. Jiang. 2012.

“Social Network

Research in Organizational Contexts: A Systematic Review
of

Methodological

Issues

and

Choices.

” Journal of

Management 38 (4): 1328

–1361.

Cope, M., and S. Elwood. 2009. Qualitative GIS: A Mixed

Methods Approach. Los Angeles, CA: Sage.

Couclelis, H. 1999.

“Space, Time, Geography.” Geographical

Information Systems 1: 29

–38.

Couclelis, H. 2010.

“Ontologies of Geographic Information.”

International Journal of Geographical Information Science
24 (12): 1785

–1809.

Galton, A. 2003.

“Desiderata for a Spatio-temporal Geo-ontology.”

In

Spatial

Information

Theory.

Foundations

of

Geographic Information Science, edited by W. Kuhn, M.
Worboys, and S. Timpf, 1

–12, Vol. 2825. Berlin: Springer.

Gatalsky, P., N. Andrienko, and G. Andrienko. 2004.

“Interactive

Analysis of Event Data Using Space-Time Cube.

” In

Proceedings

of

Eighth

International

Conference

on

Information Visualisation (IV), London. Los Alamitos, CA:
IEEE Computer Society.

Goodchild, M. F., M. Yuan, and T. Cova. 2007.

“Towards a

General Theory of Geographic Representation in GIS.

”

International Journal of Geographical Information Science
21 (3): 239

–260.

Hägerstrand, T. 1970.

“What about People in Regional Science?”

Papers in Regional Science 24 (1): 7

–24.

Janelle, D. G. 1995.

“Metropolitan Expansion, Telecommuting,

and

Transportation.

” In The Geography of Urban

Transportation, edited by S. Hanson, 407

–434. New York,

NY: The Guilford Press.

Janiak, A., ed. 2012. Kant

’s Views on Space and Time. The Stanford

Encyclopedia of Philosophy (Winter 2012 Edition), The
Metaphysics Research Lab, Center for the Study of Language
and Information. Stanford, CA: Stanford University.

Janowicz,

K.

2012.

“Observation-Driven Geo-Ontology

Engineering.

” Transactions in GIS 16 (3): 351–374.

Kristensson, P. O., N. Dahlback, D. Anundi, M. Bjornstad, H.

Gillberg, J. Haraldsson, I. Martensson, M. Nordvall, and J.
Stahl.

2009.

“An Evaluation of Space Time Cube

Representation

of

Spatiotemporal

Patterns.

”

IEEE

Transactions on Visualization and Computer Graphics 15
(4): 696

–702.

Kwan, M.-P. 2004.

“GIS Methods in Time-Geographic Research:

Geocomputation and Geovisualization of Human Activity
Patterns.

” Geografiska Annaler Series B: Human

Geography 86 (4): 267

–280.

Kwan, M.-P. 2007.

“Affecting Geospatial Technologies: Toward

a

Feminist

Politics

of

Emotion.

” The Professional

Geographer 59 (1): 22

–34.

Li, X., J. Han, and S. Kim. 2006.

“Motion-Alert: Automatic

Anomaly Detection in Massive Moving Objects.

” In

Intelligence

and

Security

Informatics,

edited

by

S.

Mehrotra, D. Zeng, H. Chen, B. Thuraisingham, and F.-Y.
Wang, 166

–177, Vol. 3975. Berlin: Springer.

Li, Z., M. Ji, J. G. Lee, L. A. Tang, Y. Yu, and J. Han. 2011.

“MoveMine: Mining Moving Object Data for Discovery of
Animal

Movement

Patterns.

” ACM Transactions on

Intelligent Systems and Technology 2 (4): 1

–32.

Nakaya, T. 2013.

“Analytical Data Transformations in Space–

Time Region: Three Stories of Space

–Time Cube.” Annals of

the

Association

of

American

Geographers

103

(5):

1100

–1106.

Palma, A. T., V. Bogorny, B. Kuijpers, and L. O. Alvares.

2008.

“A Clustering-Based Approach for Discovering

Interesting Places in Trajectories.

” In Proceedings of the

2008 ACM Symposium on Applied Computing, 863

–868.

Ceara: ACM.

Sheppard, E. 2005.

“Knowledge Production through Critical

GIS: Genealogy and

Prospects.

” Cartographica: The

International Journal for Geographic Information and
Geovisualization 40 (4): 5

–21.

Windhager, F., and E. Mayr. 2012.

“Cultural Heritage Cube. A

Conceptual Framework for Visual Exhibition Exploration.

” In

16th International Conference on Information Visualisation
(IV), Montpellier. Los Alamitos, CA: IEEE Computer Society.

Yuan, M., and A. Nara. (Forthcoming)

“Space-time Analytics

of Tracks for the Understanding of Patterns of Life.

” In

Space-Time Integration in Geography and GIScience:
Research Frontiers in the US and China, edited by M.-P.
Kwan, D. Richardson, D. Wang, and C. Zhou. Dordrecht:
Springer.

Annals of GIS

9

Downloaded by [University of Oklahoma Libraries], [Dr May Yuan] at 11:49 16 April 2014

Downloaded by [University of Oklahoma Libraries], [Dr May Yuan] at 11:49 16 April 2014