Frontiers of Architectural Research (2017) 6, 431–441

Available online at www.sciencedirect.com

Frontiers of Architectural Research

www.keaipublishing.com/foar

RESEARCH ARTICLE

Measuring the spatial and social

characteristics of the architectural
plans of aged care facilities
Ju Hyun Leea,n, Michael J. Ostwalda, Hyunsoo Leeb

a
School of Architecture and Built Environment, The University of Newcastle, Callaghan, NSW 2308,
Australia
b
Department of Interior Architecture and Built Environment, Yonsei University, Seoul 03722, Republic of
Korea

Received 26 April 2017; received in revised form 18 July 2017; accepted 13 September 2017

KEYWORDS Abstract
Visibility graph analy- This paper presents a method that combines visibility graph and isovist analyses to investigate
sis (VGA); the spatial and social properties of architectural plans for aged care facilities. The potential of
Space Syntax; the combined method is examined by measuring the properties of three sets of plans for
Aged care facility; residential aged care facilities. The first set is a pair of hypothetical, idealized plans, which
Isovist property;
allegedly reflect the “best practice” in the industry. The second set comprises a pair of plans for
Sociocultural
recent Australian designs, and the third set is a pair of plans for South Korean facilities. Results
characteristics
of the computational analysis of these six plans suggest that social and cultural factors may
shape the design of aged care settings and partially explain their international differences. The
application of this methodological approach contributes to the understanding of the relation-
ship between spaces and their cultural and social properties in the design of aged care
facilities.
& 2017 Higher Education Press Limited Company. Production and hosting by Elsevier B.V. on
behalf of KeAi. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

Winston Churchill's aphorism, “we shape our buildings and

n
Corresponding author.
E-mail address: juhyun.lee@newcastle.edu.au (J.H. Lee).
Peer review under responsibility of Southeast University.
thereafter they shape us,” has become famous in spatial
psychology, architecture, and planning. Multiple theories
have attempted to explain the reciprocal relationship

http://dx.doi.org/10.1016/j.foar.2017.09.003
2095-2635/& 2017 Higher Education Press Limited Company. Production and hosting by Elsevier B.V. on behalf of KeAi. This is an open access
article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
432
between spatial properties and human intuitive responses.
For example, Appleton's habitat and prospect–refuge the-
ories (Appleton, 1975) are among the best-known explana-
tions for environmental preference in architecture. Gibson's
ground theory (Gibson, 1950, 1979) and Kaplan and Kaplan's
information theory (Kaplan and Kaplan, 1989) are also used
to explain human responses to space. Such theories use
psychological and philosophical constructs to analyze envir-
onments and their social and behavioral properties; thus,
their results are not always reproducible (Stamps, 2005a).
Quantitative theories and methods have also been devel-
oped for understanding and modeling the relationship
between space and social patterns. Among the most famous
of these methods, in architecture at least, is Space Syntax,
which uses the mathematics of graph theory to measure
various properties of plans. Isovist analysis is a related
technique that has been traced to the work of Tandy (Tandy,
1967). An isovist is “the set of all points visible from a given
vantage point in space” (Benedikt, 1979). The mathematical
properties of isovists have potential correlations with
human spatial perceptions and responses, and they have
been used to test or apply enclosure and permeability
theories (Stamps, 2005a, 2005b; Turner et al., 2001). Space
Syntax approaches of this type have been developed for
predicting patterns of pedestrian movement, improving
surveillance, reducing crime through observation, and
understanding access and control issues well (Hillier, 1999;
Hillier and Hanson, 1984). These common computational
approaches to visibility analysis in architectural plans are
combined in visibility graph analysis (VGA).
Although these theories are useful in analyzing various
aspects of architectural planning, their application tends to
assume that human responses are relatively universal (or at
least that mathematical patterns can be derived from them,
which reflect universal tendencies). Although statistical
evidence supports the general use of several computational
methods for modeling and understanding the socio-spatial
and spatio-visual properties of plans, relatively few examples
show attempts to accommodate the needs of particular
demographic groups, such as the aged. This lacuna is
significant because recent research suggests that the spa-
tio-visual properties of physical environments have an impact
on various health-related outcomes (Hadi and Zimring, 2016;
Hendrich et al., 2009; Pachilova and Sailer, 2013; Seo et al.,
2011). Thus, researchers have used Space Syntax methods to
support the analysis of health facilities, including nurses’
behaviors (walking patterns, entries to patient rooms, and
spatial positioning), patients’ preferences and satisfaction
levels (preference for bed privacy, perceived quality of care),
and visitor movement in hospitals (Haq and Luo, 2012). For
example, Seo et al. (2011) identified that the spatial
characteristics of routes taken by nurses from intensive care
units to patient rooms and medication areas can contribute
to behavioral patterns. Setola et al. (2013) investigated the
role of spatial layout in hospitals and analyzed the integra-
tion of public and staff spaces. They revealed the patterns of
relationships between patients and the medical staff (P–M)
through the density of interactions and accessibility maps. By
combining VGA and axial map analysis, they began to develop
an understanding of the cultural differences in nursing unit
designs (Cai and Zimring, 2013). Indeed, the investigation of
hospital planning and the correlation of computational
J.H. Lee et al.
measures for spatial visibility and accessibility with beha-
vioral responses has become relatively common in recent
years (Carranza et al., 2013; Lu, 2010; Lu et al., 2009;
Morgareidge et al., 2014).
Although such studies demonstrate the usefulness of
computational analysis in health care design, they also
focus largely on functional planning and optimization. By
contrast, the challenges of designing for residential aged
care environments are not just functional but social and
cognitive. For example, past research identified that some
of the most challenging factors in the design of retirement
villages and aged care centers include isolation, loneliness,
and confusion (Gardner, 1994). Thus, the problems of
designing for aged care are not only about surveillance
and safety. The way a space is designed directly impacts the
way people socialize, stay visually connected, avoid becom-
ing lost, and become available for visits by relatives and
watching by nurses for care and safety. Aged care facilities
should be designed to accommodate social and cognitive
needs in parallel with functional requirements.
In response to this situation, this paper presents a
computational and mathematical method for capturing
selected spatial and social properties of aged care facilities.
This study revisits VGA and isovist analysis to investigate the
properties of residential aged care facilities. Setola et al.
(2013) demonstrated the usefulness of VGA in understanding
the cultural differences in health care environments, and
several related syntactical measures, including target visibi-
lity analysis, spatial positioning tool, weighted isovist area,
multilayered network, Place Syntax, and team-base and peer
distances, have been found effective in health care environ-
ments (Sadek and Shepley, 2016). In the present study, VGA is
used to develop holistic measures for examining six plans for
residential aged care facilities, and then individual isovist
measures are used to analyze differences among typical
spaces for the residents, visitors, and nurses in each facility.
To examine this combined method and its uses, the
present study analyzes six plans for residential aged care
facilities. The study commences with an overview of the
method and its application in the analysis of the plans.
Subsequently, two stages of a case study are presented,
highlighting visibility graphs and isovist properties. The
quantitative analysis of the six case study plans supports
the qualitative discussion regarding the relationship
between spaces and the cultural and social properties of
each plan. This paper concludes with a discussion on the
application of VGA and the need for future research.
2. Method and application
2.1. Visibility graph analysis (VGA)
Since the 1970s, techniques have been developed for
analyzing the 2D spatial properties of architectural plans
and the spatio-visual properties of locations within these
plans. In such techniques, the properties of an architectural
plan are typically abstracted into a set of nodes and edges
in a graph, which can then be mathematically analyzed.
This method can be used to reveal the relationship between
spaces and the social properties that necessitate or sustain
these relations (Hillier and Hanson, 1984). Convex and axial
Measuring the spatial and social characteristics of the architectural plans of aged care facilities
Figure 1 Isovist viewshed (grey-shaded polygon) from a location in a plan identifying several commonly used measurable properties.
maps are two of the most common approaches to spatial
analysis using graph theory. Convex maps allow for the
analysis of enclosed or defined programmatic spaces and
their connectivity, which are conceptualized as nodes and
edges, respectively. Environmental and behavioral research
defines enclosure as an important property of a space
because it limits or shapes movement and perception
(Alitajer and Molavi Nojoumi, 2016; Hayward and Franklin,
1974; Nasar and Jones, 1997; Stamps, 2005a). By contrast,
an axial map is defined as the set of the longest lines of sight
or movement that pass through and connect all the habi-
table spaces of a plan (Hillier and Hanson, 1984). Axial maps
capture behavioral characteristics, including movement
potential and navigational choice, which are represented
by edges and nodes, respectively. In addition to axial and
convex maps, a third approach constructs a series of isovist
polygons around a regular grid in space and then compares
the connectivity relationships between them. This isovist
“field” is a representation of the spatio-visual properties of
an environment. The properties of the isovist's geometry or
construction are shown in Figure 1.
Regardless of which approach among the three is
adopted, graph mathematics is then used to develop
quantitative mathematical descriptions of the features of
the built environment that are represented by edges and
nodes. The mathematical measures derived from the maps
include connectivity (number of direct connections to other
spaces or movement paths), integration (comparison of
normalized mean depth to an optimal D-value), and intel-
ligibility (correlation between connectivity and integra-
tion). Connectivity describes the mutual relationship
between all movement paths or spaces, whereas integration
indicates how well one space or path is connected to all
others in a system (Hillier and Hanson, 1984). The measures
derived from the isovist field for a plan effectively consti-
tute the visibility graph of its properties. Past research has
suggested that integration measures, including local and
global integration, can be used to shape the design of health
care facilities. “Visibility structures work together with and
enhance permeability structures,” and both influence how
an environment functions spatially and how it is experi-
enced and understood (Guney, 2007).
433
The present study uses two of these methods, one for
each stage of the analytical process. Although both methods
are connected and related through the isovist field, they
have different levels of empirical evidence and traditional
applications (Braaksma and Cook, 1980; Turner et al., 2001).
Although syntactical and isovist approaches have various
levels of empirical support, VGA seems to offer a potential
for developing holistic insights into the design of buildings
that have visibility requirements. By contrast, isovists are
effective in analyzing spatial characteristics in terms of the
perspectives of user groups.
In the first stage of the analysis, VGA is used to measure
the connectivity and visual integration of the six plans for
residential aged care facilities. The isovist field is generated
for all plans using a consistent 500 mm grid, and all VGA
calculations are completed using UCL Depthmap software
(Turner, 2001). The mean results for each plan for isovist
area and perimeter, drift angle and magnitude, maximum
and minimum radial, and occlusivity are then derived.
These measures provide a holistic indication of the various
spatial properties of the plan. These measures have also
been mapped to various cognitive and behavioral aspects of
architectural plans (Dawes and Ostwald, 2014a, 2014b).
2.2. Isovist analysis
In the second stage of the analysis, specific isovists within
each plan are used to examine the properties of different
spaces and user experiences. For this purpose, isovist area
and perimeter measures are used to investigate four
selected locations in each plan. The four locations are as
follows: lobby (the “visitor” perspective), common area
(dining or lounge room, representing the “communal”
perspective), nurse room (the “carer” perspective), and
corridor in residential units (the “resident” perspective).
After developing the four isovists of each case, four proper-
ties, namely, area, perimeter, max radial, and occlusivity,
are measured and analyzed for each location. Figure 2
illustrates a hypothetical case with four vantage points and
their isovists.
434
Figure 2 Hypothetical case with four vantage points generated from representative test locations and their isovists.
The four isovist measures are selected after a review of
the results of past research, which examined the mathema-
tical properties of isovists with human spatial perceptions
(Stamps, 2005a). For example, Dawes and Ostwald (2014a,
2014b) addressed five isovist properties (i.e., isovist area,
radial line, length, occlusion, and drift) that possess a level
of correlation with empirical research. Recent research has
also demonstrated the statistically significant correlation
between measures and perceptions of spatial exposure and
enclosure (Dosen and Ostwald, 2017). The different spatio-
visual analytical approaches used in architecture also corre-
late closely with the major models of spatial cognition (Franz
et al., 2005).
In practical terms, VGA and isovist methods are generally
applied to 2D representations of space (typically plans, but
sometimes sections), and the correlation data for human
perceptions are similarly derived from such views. In the
present study, the plans of the six designs are produced, and
the analysis focuses only on interior spaces and views.
Furthermore, doors, low partitions, and furniture below
eye level are removed from each plan prior to conducting
the VGA. Therefore, the results represent the inter-visibility
properties of each layout (Cai and Zimring, 2013).
2.3. Cases
The floor plans of three sets of designs, (i) idealized, (ii)
Australian, and (iii) Korean aged care settings, are investi-
gated in this study. The first set, which comprises two plans,
are hypothetical layouts for spatial environments that are
perceived as ideal for residential aged care facilities. Both
plans were developed in the United States of America, the
first in 2004 and the second in 2013. The next pair of plans
are for Australian residential aged care facilities and were
J.H. Lee et al.
constructed in New South Wales in 2012. The third set of
plans is for South Korean residential aged care facilities, one
from 2004 and the other from 2015, both in Seoul. Although
these six cases cannot represent the architectural design
approaches of the aged care facilities in each nation, they
can be regarded as contemporary designs that are shaped by
sociocultural characteristics and values.
The first set of the two cases is selected from past
research (Brawley, 2006; Eastman, 2013). One is derived
from Eastman's ideal plan for an assisted living and group
residence, which is denoted as “H1.” The other is from
Brawley's building configuration typologies for ideal resi-
dences for aged care, which is denoted as “H2.” H1 has a
broadly H- or X-shaped plan, whereas H2 conforms to a
“race track” typology. The selected cases were chosen, in
part, because they feature double- and single-loaded
corridors, which are common in aged care facilities.
However, despite these general differences, specific pro-
gram requirements, budget realities, site considerations,
building and planning regulations, and many other factors
shape the actual layout of a plan. Nevertheless, the two
hypothetical plans demonstrate an ideal zoning or layout
for an aged care facility.
Across the last two sets, the number of units ranges from
86 to 307 (Table 1). The comparative study in the next
section uses the mean values of measures to allow for useful
comparisons given the scale differences. Australian aged
care residences often adopt double-loaded corridors (a
room on each side of a central corridor), whereas the two
Korean examples chosen feature single-loaded corridors
despite the relatively high cost (construction and land costs)
and inefficiency associated with this planning strategy.
The two Australian facilities (Au1 and Au2) are multi-
level, that is, two- or three-story buildings with the same
Measuring the spatial and social characteristics of the architectural plans of aged care facilities 435

Table 1 Four aged care facilities in Australia and South Korea.

Nation Australia South Korea

Code Au1 Au2 Ko1 Ko2

Name Nambucca Head Uniting- Port Macquarie Uniting- Songpa Aging Skilled- Gangnam Happiness Nur-
Care Aging Care Aging nursing home sing hospital
Date completed January 2012 February 2012 March 2004 January 2015
No. of care units 86 110 93 307
Location Nambucca Heads, NSW Port Macquarie, NSW Songpa-gu, Seoul Gangnam-gu, Seoul
Site size 3853 m2 14,677 m2 1500 m2 6426 m2
2
Building footprint 1921 m 3635 m2 407 m2 3150
Building area 5463 m2 7225 m2 3116 m2 18,585 m2

Figure 3 Visibility graphs with connectivity values (Con.) of Figure 4 Visibility graphs with connectivity values (Con.) of
hypothetical cases. Australian cases.

example, Figs. 3–5 contain the visibility graphs for each

planning on each floor. Thus, only one level of each case
should be examined. The two South Korean facilities (Ko1
and Ko2) are five and six floors high. The ground floors of the
two Korean facilities differ from the upper levels, with the
former containing extensive administration areas and a
café. Thus, we examine the ground floor and a standard
residential floor for Korean cases in the following sections.
3. Case study
3.1. Stage 1
The VGA process maps various values (color coded from red
for “very high” measures to dark blue for “very low”
measures for representation) to each grid square. For
case, which are colored to represent connectivity. Although
each cell has a connectivity value assigned to it for detailed
numerical analysis, color coding allows for the development
of an early intuitive reading.
Figure 3 shows that the four wings in H1 have low
connectivity values (dark blue), thereby becoming visually
and socially isolated in the plan. In contrast, H2 has a
relatively small volume of space with the same connectivity
value, being mostly bathrooms and service areas. There-
fore, the plan of H1 is more isolating and hierarchical than
that of H2, whereas the plan of the latter is more inclusive
and accommodating of social interaction than H1. The
connectivity ranges or values for the Australian cases in
Figure 4 are similar although the design with the larger
footprint (Au2) has a lower level of connectivity compared
with those of the other design. By contrast, Ko2 in the
436 J.H. Lee et al.

circulation spaces and rely on single-loaded corridors. In

general, a high integration value represents a shallow or an
assessable space (Hillier and Hanson, 1984). Although the
hypothetical plans of set one have no site constraints, the
“double-loaded corridor” (H1) and “race track” (H2) typol-
ogies have a middle level of integration among the cases.
The two Australian cases have relatively weak or low
connectivity and integration measures. The results are
unexpected because Australia has extensive experience in
the aged care domain, and a high mean integration level is
anticipated. The double-loaded corridors, interpenetration
of landscaped courtyards, and significant reliance on hor-
izontal (rather than vertical) circulation shape this result.
Notably, H1 and Ko2 have mean isovist area values that are
more than double those in Au1 and Au2 (84.74 and 147.97,
86.07 4 37.98, and 35.17) and a drift magnitude (often
described as “strength of enticement” or “visual pull”) that
is of a similar level of difference. However, Ko2's occlusivity
is also the highest among the cases, thereby indicating that
it possesses a high level of mystery or presents an oppor-
Figure 5 Visibility graphs with connectivity values (Con.) of
tunity to hide or become lost.
Korean cases.

Korean set has the highest level of connectivity and has
more extensive areas of warm colors (reds and oranges) in
their graphs than Ko1. Across the six plans, lobbies and
corridors have higher connectivity than the other locations.
However, the actual measures should be assessed to inter-
pret the results thoroughly (Table 2).
The visual integration values for Ko1 and Ko2 (VI = 4.58
and 5.48) are higher than those of the hypothetical (VI =
3.79 and 4.36) and Australian sets (VI = 2.25 and 3.01)
possibly because multistory Korean facilities have focused
3.2. Stage 2
Stage 2 analyzes isovists in four distinct locations on each
plan, which represent three different users (residents,
visitors, and nurses) and a general communal space. The
results for H1 (Table 3), as a “control” plan, feature a clear
hierarchy wherein Lobby 4 Common 4 Corridor / Nurse.
Visitors have the strongest or most distinctive spatial
experience, having the highest levels of visible space (out-
look), distance from walls (IMin), distance to furthest view

Table 2 Mean values of spatial properties (VI: visual integration, IA: isovist area, IP: isovist perimeter, IDA: isovist drift
angle, IDM: isovist drift magnitude, IMax: isovist max radial, IMin: isovist min radial, IO: isovist occlusivity, GF: ground floor,
and RF: residential floor).

VI IA (m2) IP (m) IDA IDM (m) IMax (m) IMin (m) IO (m)

H1 3.79 84.74 76.62 180.09 4.10 15.56 1.00 38.69

H2 4.36 55.41 65.29 181.38 4.36 16.71 0.69 31.97
Au1 3.01 37.98 50.48 181.20 2.97 11.79 0.61 25.00
Au2 2.51 35.17 47.60 177.92 2.86 11.39 0.59 22.68
Ko1 - GF 7.11 57.11 66.27 175.63 3.51 12.12 0.87 36.88
- RF 5.48 44.79 64.59 170.25 3.51 12.82 0.63 37.13
Ko2 - GF 5.01 147.97 127.03 183.67 5.71 22.13 1.05 78.06
- RF 4.58 86.07 92.52 178.95 4.65 19.48 0.83 54.73
Mean 4.48 68.66 73.80 178.64 3.96 15.25 0.78 40.64
SD 1.45 37.40 25.73 4.17 0.95 3.96 0.18 18.01

Table 3 Isovist properties from four vantage points in H1.

IA (m2) IP (m) IDM (m) IMax (m) IMin (m) IO (m)

Lobby 243.72 145.01 4.13 25.23 4.80 71.56

Common 110.91 99.00 3.04 13.83 2.65 55.83
Nurse 30.95 50.63 1.17 10.02 1.62 26.51
Corridor 32.85 63.86 0.45 12.04 0.58 24.64
Measuring the spatial and social characteristics of the architectural plans of aged care facilities 437

Table 4 Isovist properties from four vantage points in the Australian and Korean cases (IAP: isovist area proportion = [IA/
Building area]  100).

IA (m2) IAP (%) IP (m) IDM (m) IMax (m) IO (m)

Au1 Lobby 81.21 1.49% 81.05 13.15 1.67 34.51

Common 68.09 1.25% 51.47 10.29 0.57 17.21
Nurse 59.93 1.10% 75.58 21.40 1.13 44.60
Corridor 54.04 0.99% 69.89 14.81 0.89 29.24
Au2 Lobby 73.55 1.02% 79.22 0.81 8.94 43.09
Common 58.63 0.81% 43.63 0.44 7.78 11.07
Nurse 21.98 0.30% 41.37 1.68 9.01 22.65
Corridor 74.10 1.03% 122.76 0.60 19.30 65.53
Ko1 Lobby 98.79 3.17% 112.94 0.73 10.09 68.03
Common 59.44 1.91% 61.58 3.24 14.40 30.44
Nurse 80.40 2.58% 101.27 3.24 11.08 65.07
Corridor 72.53 2.33% 132.23 0.51 13.66 87.01
Ko2 Lobby 384.31 2.07% 303.48 9.02 33.62 217.07
Common 124.82 0.67% 150.67 1.89 21.43 99.70
Nurse 173.21 0.93% 222.40 0.80 21.83 146.28
Corridor 183.98 0.99% 180.43 1.83 28.38 110.14
Mean 104.31 1.42% 114.37 5.28 12.74 68.23
SD 85.87 0.78% 71.45 6.46 9.98 54.52

limits (Imax), and occlusivity. Ironically, they experience
the greatest levels of exposure and mystery. The common
area demonstrates middle-level visibility in terms of the
isovist viewshed.
The isovist properties from four vantage points in the
Australian and Korean cases are measured in Table 4, and
their isovist maps are depicted in Figs. 6 and 7. The isovist
area can be affected by the building area, and the plans
analysed in this paper differ in size. Thus, the proportion of
the total area that is occupied by a given isovist is also
calculated, that is, the isovist area proportion (IAP). Lobbies
in both Korean cases have higher area and occlusivity than
those in the Australian cases. The nurse room and residen-
tial corridor in Ko1 have relatively high IAP because it has
the smallest building footprint of the six plans even though
it has a “group home” layout with only south-facing resident
units, which resembles the traditional Korean housing
configuration.
To effectively conduct a comparative analysis, four
properties—IAP, perimeter (IP), max radial (IMax), and
occlusivity (IO)—are charted and compared for each loca-
tion (Figure 8). When charted, these results reinforce
several of the interpretations identified in the holistic
examination of the plan and point to some patterns and
differences, which are seen in the chart for IAP, thus
confirming that the Korean cases do feature a large volume
of visible space in the visitor's lobby but a small volume in
the common area. However, the ISPs of the Australian
cases show a similar pattern to H1, which features a
hierarchical pattern from visitor to nurse and resident,
thereby possibly implying that the Australian cases are
closer than the other plans to the ideal building layout for
the “double-loaded corridor” type, which is a traditional
western typology for aged care facilities. The different
isovist properties (Table 4 and Figure 8) may also reflect
the differences between these spatial configurations in
cultural terms.
Interestingly, Au1's drift magnitude indicates a dramatic
visual pull in the nurse room and the residential corridor,
whereas Au1's max radial values are relatively short and
unchanged between locations. This spatial property illumi-
nates the contemporary idea that the nursing station has a
lower level of occlusivity (hidden spaces) than any other
area, whereas the visitor's lobby appears to have the lowest
level of occlusivity in all cases (being more exposed). This
observation may be because the design of aged care facilities
highlights the clear relationship between visitors and resi-
dents, which is different from that found in a common
hospital layout. However, it also suggests that visibility
should be clear for the nurses in aged care facilities.
4. Discussion
This paper describes and demonstrates a computational
approach for measuring and extrapolating the spatial and
social properties of residential aged care facilities. This
method can capture the syntactical and visibility character-
istics of the architectural plans, which may be examined
quantitatively and qualitatively. The method is demonstrated
by analyzing six aged care facilities. Acknowledging the small
sample size, the data derived from the demonstration are
used to provide an insight into the spatial and social patterns
of each case, which have similarities and disparities.
The results indicate that the hypothetical and Australian
cases have a clear spatial hierarchy from the lobby to the
corridor or nursing station (from visitor to resident), the
Korean cases possess quite different patterns (Figure 8).
These spatial properties in each case may be related to the
design typologies that architects have adopted, either from
438
Figure 6 Isovist maps from four selected points in the Australian cases.
policy guides or from a priori knowledge or reasoning. The
research on the design of residential aged care facilities has
noted a gradual change in the way they are planned in
response to the increased understanding of the challenges
involved (Brawley, 2006). For example, the traditional
“double-loaded corridor” model can evolve into a “race
track” design or a “mall” configuration. A “group home”
layout is often expected to be planned like a “cluster of
group homes” (Brawley, 2006). Thus, comparative case
study methods should be able to measure and differentiate
the properties of building configuration typologies. For
example, the Australian cases use a “double-loaded corri-
dor” configuration, which is a standard western planning
tradition, whereas the Korean cases attempt to adopt a new
version of building configuration typologies that are close to
the country's historic types. Thus, the first Korean case
J.H. Lee et al.
features a multistory “group home” type, whereas the
second Korean case is a combination of the mall and
courtyard types.
The results of the isovist analysis from four locations also
indicate that social and cultural differences exist among the
visibility configurations of the different sets of facilities. For
example, Korea's aging population is likely to be familiar
with public housing and apartments even though they may
prefer to live with their first child or “age in place” (in their
own home). By contrast, Australia's aging population might
be accustomed to free-standing homes (the “great Austra-
lian dream”), with community facilities and social support.
Thus, the plans (and isovist properties of common areas)
may reflect these social and cultural differences. That is,
the results of VGA and isovist analysis in this study imply
that syntactical and cultural characteristics can be observed
Measuring the spatial and social characteristics of the architectural plans of aged care facilities
Figure 7 Isovist maps from four selected points of the Korean cases.
in the aged care settings, thus reflecting the residents’
sociocultural preferences or needs. The differences may be
caused by the architects’ intentions—for example, Korean
cases are designed for improving the self-esteem of the
elderly, whereas the Australian cases focus on community
and affordable care—but one's strength can be used to
overcome the weakness of the other designs.
5. Conclusion
This study conducted a computational analysis of building
plans designed for residential aged care facilities. The
paper presents a combined method for measuring the
syntactical values and isovist properties of a case study
set of three pairs of plans: hypothetical, Australian, and
Korean aged care facilities. All tests were performed using a
combination of architectural plans, geometric data, and
software. The results of the analysis indicate that the
proposed method captures the spatial and social character-
istics of the design of aged care facilities. Although archi-
tectural design decisions result from a variety of contextual
439
factors, such as local regulations, budget realities, site
considerations, planning codes, and construction regula-
tions, our method reveals common and specific design
patterns in plan layouts. Thus, the method can also be
applied to investigate the environmental and behavioral
properties in other building types.
The results presented in this paper also indicate the
viability of using these measures to differentiate between
spaces in terms of their visibility properties in aged care
settings. For example, the two Korean designs in this study
have the highest integration values and the largest isovist
areas and perimeters, thus showing that they are both well
connected spatially and have a larger volume of visible
space from most positions on different floors than the other
designs. The multistory designs (more than five floors) also
have interesting visibility configurations. Despite the lim-
ited dataset, this finding is a positive result, but it can
reflect an institutional approach to planning. The authors
plan a future study that addresses the relationship between
spatial configuration and social cognition with more varied
cases and cultural differences than in the current study.
440
Figure 8 Isovist properties from four selected points in Australian and Korean cases.
References
Alitajer, S., Molavi Nojoumi, G., 2016. Privacy at home: analysis of
behavioral patterns in the spatial configuration of traditional
and modern houses in the city of Hamedan based on the notion
of space syntax. Front. Archit. Res. 5, 341–352.
Appleton, J., 1975. The Experience of Landscape (reprinted in
1996). Wiley, New York.
Benedikt, M.L., 1979. To take hold of space: isovists and isovist
fields. Environ. Plan. B: Plan. Des. 6, 47–65.
Braaksma, J.P., Cook, W.J., 1980. Human orientation in transporta-
tion terminals. Transp. Eng. J. Am. Soc. Civil. Eng. 106, 189–203.
Brawley, E.C., 2006. Design Innovations for Aging and Alzheimer's:
Creating Caring Environments. John Wiley & Sons, Hoboken, NJ.
Cai, H., Zimring, C., 2013. Understanding cultural differences in
nursing unit design with the support of space syntax analysis: are
Chinese nursing units designs different from their U.S. counter-
parts?. In: Kim, Y.O., Park, H.T., Seo, K.W. (Eds.), Proceedings of
the Ninth International Space Syntax Symposium. Sejong Uni-
versity, Seoul, pp. 1–24.
Carranza, P.M., Koch, D., Izaki, A., 2013. SPOT with paths, and
interactive diagram with a low complexity isovist algorithm. In:
Kim, Y.O., Park, H.T., Seo, K.W. (Eds.), Proceedings of the Ninth
International Space Syntax Symposium.. Sejong University,
Seoul, pp. 1–13.
Dawes, M.J., Ostwald, M.J., 2014a. Prospect-Refuge theory and the
textile-block houses of Frank Lloyd Wright: an analysis of spatio-
visual characteristics using isovists. Build. Environ. 80, 228–240.
J.H. Lee et al.
Dawes, M.J., Ostwald, M.J., 2014b. Testing the ‘Wright Space’:
using isovists to analyse prospect-refuge characteristics in
Usonian architecture. J. Archit. 19, 645–666.
Dosen, A.S., Ostwald, M.J., 2017. Lived space and geometric space:
comparing people's perceptions of spatial enclosure and expo-
sure with metric room properties and isovist measures. Archit.
Sci. Rev. 60, 62–77.
Eastman, P., 2013. Building Type Basics for Senior Living. John Wiley
& Sons, Hoboken, NJ.
Franz, G., Mallot, H., Wiener, J., Neurowissenschaft, K., 2005.
Graph-based models of space in architecture and cognitive
science-a comparative analysis. In: Proceedings of the 17th
International Conference on Systems Research, Informatics and
Cybernetics. Baden-Baden, pp. 30–38.
Gardner, I.L., 1994. Why people move to retirement villages: home
owners and non-home owners. Aust. J. Ageing 13, 36–40.
Gibson, J.J., 1950. The Perception of the Visual World. Riverside
Press, Cambridge.
Gibson, J.J., 1979. The Ecological Approach to Visual Perception
(reprinted in 1986). Lawrence Erlbaum Associates, Hillsdale, NJ.
Guney, Y.I., 2007. Analyzing visibility structures in Turkish domestic
spaces. In: Proceedings of 6th International Space Syntax
Symposium. İstanbul, pp. 1–12.
Hadi, K., Zimring, C., 2016. Design to improve visibility: impact of
corridor width and unit shape. Health Environ. Res. Des. J. 9, 35–49.
Haq, S., Luo, Y., 2012. Space syntax in healthcare facilities
research: a review. Health Environ. Res. Des. J. 5, 98–117.
Hayward, S.C., Franklin, S.S., 1974. Perceived openness-enclosure
of architectural space. Environ. Behav. 6, 37–52.
Measuring the spatial and social characteristics of the architectural plans of aged care facilities
Hendrich, A., Chow, M.P., Bafna, S., Choudhary, R., Heo, Y.,
Skierczynski, B.A., 2009. Unit-related factors that affect nursing
time with patients: spatial analysis of the time and motion
study. Health Environ. Res. Des. J. 2, 5–20.
Hillier, B., 1999. Space is the Machine: a Configurational Theory of
Architecture. Cambridge University Press, Cambridge.
Hillier, B., Hanson, J., 1984. The Social Logic of Space. Cambridge
University Press, Cambridge.
Kaplan, R., Kaplan, S., 1989. The Experience of Nature: a Psycho-
logical Perspective. Cambridge University Press, Cambridge.
Lu, Y., 2010. Measuring the structure of visual fields in nursing units.
Health Environ. Res. Des. J. 3, 48–59.
Lu, Y., Peponis, J., Zimring, C., 2009. Targeted visibility analysis in
buildings: correlating targeted visibility analysis with distribu-
tion of people and their interactions within an intensive care
unit. In: Koch, D., Marcus, L., Steen, J. (Eds.), Proceedings of
the Seventh International Space Syntax Symposium.. Sejong
University, Seoul, pp. 1–10.
Morgareidge, D., Cai, H., Jia, J., 2014. Performance-driven design
with the support of digital tools: applying discrete event
simulation and space syntax on the design of the emergency
department. Front. Archit. Res. 3, 250–264.
Nasar, J.L., Jones, K.M., 1997. Landscapes of fear and stress.
Environ. Behav. 29, 291–323.
Pachilova, R., Sailer, K., 2013. The effect of hospital layout on
caregiver-patient communication patterns. In: Yoxall, A., Chris-
ter, K. (Eds.), Proceedings of the Second European Conference
441
on Design 4 Health. Sheffield Hallam University, Sheffield, UK,
pp. 174–184.
Sadek, A.H., Shepley, M.M., 2016. Space syntax analysis: tools for
Augmenting the precision of healthcare facility spatial analysis.
Health Environ. Res. Des. J. 10, 114–129.
Seo, H.-B., Choi, Y.-S., Zimring, C., 2011. Impact of hospital unit
design for patient-centered care on nurses' behavior. Environ.
Behav. 43, 443–468.
Setola, N., Borgianni, S., Martinez, M., Tobari, E., 2013. The role of
spatial layout of hospital public spaces in informal patient-
medical staff interface. In: Kim, Y.O., Park, H.T., Seo, K.W.
(Eds.), Proceedings of the Ninth International Space Syntax
Symposium. Sejong University, Seoul, pp. 1–11.
Stamps, A.E., 2005a. Isovists, enclosure, and permeability theory.
Environ. Plan. B: Plan. Des. 32, 735–762.
Stamps, A.E., 2005b. Visual permeability, locomotive permeability,
safety, and enclosure. Environ. Behav. 37, 587–619.
Tandy, C.R.V., 1967. The isovist method of landscape survey. In:
Murray, H.C. (Ed.), Symposium: Methods of Landscape Analysis.
Landscape Research Group, London.
Turner, A., 2001. Depthmap: a program to perform visibility graph
analysis, Proceedings 3rd International Symposium on Space Syntax.
Georgia Institute of Technology, Atlanta, GA, 31.31–31.39.
Turner, A., Doxa, M., O'Sullivan, D., Penn, A., 2001. From isovists to
visibility graphs: aa methodology for the analysis of architec-
tural space. Environ. Plan. B: Plan. Des. 28, 103–121.