Database Reference
In-Depth Information
association relationships and taking into account
imprecise measures. (Un)Fortunately, visualiza-
tion of multidimensional imprecise (spatial) data
is an open issue (Pang, 2008).
Spatial OLAP models and tools are based on
the vector representation of spatial data. Field
data, which represent spatial data as a regular grid
whose cells are associated with alphanumeric at-
tributes, could be used for spatio-multidimensional
analysis. Indeed,Ahmed & Miquel (2005) provide
a continuous representation of spatial dimensions,
but the introduction of field data as analysis subject
remains unexplored. Aggregation of field data
could be supported by MapAlgebra operators. Map
Algebra defines a set of operations on field data
(local, focal and zone operators) (Tomlin, 1990).
Adaptation of Map Algebra to multidimensional
data structures, definition of pre-aggregation and
visualization/interaction techniques for continu-
ous measures are challenges to overcome for an
effective SOLAP tool supporting field data.
The integration of trajectories data into data
warehouses raises several problems because
classical multidimensional models are based on
discrete facts and dimensions and they do not
take into account spatial predicates. This problem
has been investigated by some works (Wan et al.,
2007; Orlando et al., 2007) in the last years. How-
ever, the definition of a SOLAP client to visually
query and analyze trajectory data warehouses is
an unexplored important challenge.
Bertolotto et al. (2007) affirm that visual
analysis of spatial data mining results is improved
by exploiting the third dimension of spatial data
through an interactive 3D geovisualization system.
The integration of advanced geovisualization
techniques within OLAP clients in order to support
multidimensional 3D spatial data is an interesting
research direction.
Finally, several semiology problems have to
be solved for the correct and relevant visualiza-
tion of measures. Measures can be displayed with
labels but it is sometimes worthwhile to use more
expressive, significant visual components. The
way measures will be displayed on the map must
depend on several criteria: nature of the measure
(quantitative or qualitative measure), number
of measures to be displayed, and current repre-
sentation of the spatial dimension (point, line or
polygon). Moreover, GIS users are usually specific
knowledge domain decision makers. GIS takes
into account their profiles and preferences in order
to provide well-suited cartographic visualization
(Vangenot, 2001). Thus, the ambition is to define
a method to automatically find out the most ap-
propriate cartographic representation of SOLAP
queries results. The visual variables (size, colours,
etc.) and the graphic representation (i.e. bar, pie,
etc.) used to represent measures on maps can be
automatically deduced thanks to SOLAP query
patterns taking into account number, type and
current representation of dimensions, measures
types, aggregation functions involved in the query
and user profile and/or preferences (Bellatreche,
et al., 2005).
CONCLUSION
Spatial OLAP refers to the introduction of spatial
data into data warehouse and OLAP systems.
SOLAP enhances decision analysis capabilities
of OLAP systems allowing exploiting the com-
plex nature of geographic information. SOLAP
redefines main OLAP concepts. It defines spatial/
geographic dimensions as dimensions with spa-
tial attributes, spatial measures as a collection of
spatial objects or the result of spatial operators,
and geographic measures as geographic objects
belonging to hierarchy schemas. SOLAP extends
multidimensional navigation operators defining
spatial drill and cutting operators which allow
navigating into spatial/geographic dimensions and
cutting the hypercube thanks to spatial and non-
spatial predicates. Other spatio-multidimensional
operators permit to change the structure of the spa-
tial hypercube thanks to spatial analysis operators,
to permute dimensions and geographic measures,
Search WWH ::
Custom Search