Geoscience Reference
In-Depth Information
Package for the Social Sciences [SPSS]), Geographic Information Systems (GIS) (e.g. ArcGIS) or
modelling frameworks (e.g. Repast). For less experienced users, the workflow might be rather less
desirable. Such users might include policymakers, visitors from external academic disciplines or
early career researchers. For any or all members of this diverse group, a much more readily digest-
ible means of access to social simulation technologies would almost certainly promote their uptake
and usage. A major component of the architecture layer of the NeISS is therefore a further level of
integration which provides combinations of services in the form of a portal rather than as individual
workflows. In view of the combination of services which it entails, the portal can perhaps best be
likened to a web-based
spatial decision support system
, by making maps, models, software and
data available to users as ready-processed intelligence or even knowledge, relating to the underlying
system (cf. Geertman and Stillwell, 2003). In effect, such portals provide the means to configure
applications for the needs of specific user communities, which is effectively what we mean by the
term
virtual organisation
in our earlier discussion.
Other important activities at this stage of the process are security and access to computation. In
respect of the former, the key challenge is to regulate access to simulation services - this will be
particularly important if access involves issues relating to copyright, ethics or commercial licences
relating to data, software or models. In relation to the latter, a key feature of simulation is resource
intensity in any of the key phases of data processing, model execution or visualisation. The provision
of seamless access to the necessary resources remains an important requirement, even though it is
by no means the sole or dominant rationale for e-Research, as we have been at pains to argue earlier.
The
top
layer of the NeISS infrastructure brings the important perspective of end users into
the picture, emphasising again the fact that the e-Research perspective is concerned with a loose
coupling of technology components. Thus, users' interactions with the infrastructure are part of a
bigger life cycle picture in which the needs and experiences of users are a key part of the feedback
mechanism which stimulates the ongoing development and evolution of research capabilities.
10.4 E-RESEARCH APPLICATIONS AND USES
The most important application domain for the NeISS infrastructure is clearly to support academic
research. A key principle to note here is that this constituency is by no means limited to geographers.
Indeed, one might argue that while (some) geographers may themselves have the skills to bring
together data, analysis, models and visualisation as a means to address difficult research ques-
tions, this capability is substantially less likely within other social science disciplines which may
nevertheless share an interest in many similar problems. For example, Malleson and Birkin (2011)
have demonstrated how the infrastructure may be configured to support the integration of micro-
simulation with agent-based modelling approaches to the study of crime. A major benefit of this
approach is that it allows a balance to be struck between the roles of both criminal and victim. Thus,
engagement with new forms of data, including volunteered sources (e.g. OpenStreetMap) as well
as property gazetteers, land use data and high-resolution mapping of buildings, road networks and
accessibility surfaces, provides a rich backdrop for models of criminal opportunity and behaviour.
Teaching is another major application space of interest to the academic community. A simple
portal has been developed for students at the University of Leeds in which small area populations
of the city are projected 30 years into the future, and these simulations are then perturbed under
the influence of various scenarios for changing delivery of health care, housing and transporta-
tion. In effect, therefore, this activity combines a series of workflows as shown in Figure 10.2. In
the baseline, census data feed a microsimulation model (MSM) of small area populations, which
are then projected forward in time through a dynamic model. The dynamic model in turn feeds
a group of sector-specific spatial interaction models (SIMs), each of which is encapsulated as a
range of analytical performance indicators. The outcomes can then be generated as maps and
tables. Different scenarios can be produced by plugging in alternative distributions of
supply
(i.e. the provision of health care, housing or transport) into the simulation workflow. Although this
