Database Reference
In-Depth Information
tographic representation of geographic data. On
the contrary, maps are fundamental for the spatial
decision making process because they stimulate
user's cognitive process, and reveal hidden pre-
cious geospatial information. Therefore, some
solutions, called Spatial OLAP, which integrate
GIS cartographic visualization and interaction
functionalities into OLAP systems, have been
developed. Spatial OLAP (SOLAP) is “a visual
platform built especially to support rapid and easy
spatio-temporal analysis and exploration of data
following a multidimensional approach comprised
of aggregation levels available in cartographic
displays as well as in tabular and diagram displays”
(Bédard, 1997). Spatial OLAP systems integrate
advanced OLAP and GIS functionalities (Rivest
et al., 2005; Kouba et al., 2000). They visualize
measures on maps at different spatial granularities
revealing relations between facts and dimensions
(Bédard, et al., 2001). Moreover, maps allow trig-
gering spatio-multidimensional operators through
simple mouse clicks, also. Different SOLAP
models have been proposed. They address various
aspects of geographic information allowing to
model different spatio-multidimensional applica-
tions. SOLAP applications can address several
and different domains: environmental studies,
marketing, archaeology, epidemiology, etc.. SO-
LAP models define the concepts of spatial/geo-
graphic dimension, spatial/geographic measure,
and spatio-multidimensional operators.
The integration of spatial data into multi-
dimensional models and systems rises several
theoretical and implementation issues. Therefore,
in this chapter, we introduce main OLAP and
GIS concepts. Then, a detailed review of SO-
LAP models, architectures and research SOLAP
tools is presented. The chapter describes our
Web-based prototype for the analysis of spatio-
multidimensional databases (GeWOlap) (Bimonte
et al., 2006; Bimonte et al., 2007a, Bimonte et al.,
2007b). We describe main architectural features,
and we present spatio-multidimensional and GIS
operators using a study case concerning pollution
in French cities. Main outcome and limits of our
approach as regards to existing SOLAP tools are
detailed, also. Finally, future research directions
in spatio-multidimensional visualization and
interaction are discussed.
BACKGROUND
Data Warehouse and OLAP Systems
Data warehouse and OLAP systems are business
intelligence tools intended to support multidimen-
sional analysis of huge datasets. Data are modelled
according to the multidimensional model, which
is based on the concepts of dimensions and facts
(Inmon, 1996). Dimensions represent analysis
axes. They are organized in hierarchies' schemas.
Facts, described by numerical values (measures),
are subjects of analysis. Measures are analyzed
at different granularities corresponding to dimen-
sion hierarchies' levels, and they are aggregated
by means of SQL aggregation functions. The
instance of a multidimensional model is the hy-
percube. It is a set of cells representing measures
at all combinations of dimensions' levels. OLAP
operators permit to navigate into the hypercube.
Most common operators are drill and cut operators.
Drill operators (i.e., Roll-Up and Drill-Down) let
navigating into dimension hierarchies aggregat-
ing measures. Cut operators (i.e. Slice and Dice)
permit to reduce the analysis space, by selecting
a sub-set of dimensions members.
Usually, data warehouse and OLAP systems
are based on a three-tier architecture. The first tier
is the data warehouse, where data, coming from
external heterogeneous sources, are uniformed and
stored according to the multidimensional model.
The second tier is the OLAP Server. It imple-
ments OLAP operators and pre-computes a set of
multidimensional queries to grant effective query
response times. The OLAP server implements
other advanced functionalities also, such as control
accesses, multidimensional calculation engine,
Search WWH ::
Custom Search