Global Positioning System Reference
In-Depth Information
Load district shape fi le
Establ ish SRS
Verify dis rict code
t
Select req uired data
Save wrong data
Sort by district code
Group
districts
Verify geometry
Calculate area
Establish county
Store districts
Correct geometry
Fig. 6. Transformations applied to the geometries of the District shape fi le.
Color image of this figure appears in the color plate section at the end of the topic.
Creation of SOLAP cubes
A typical architecture of DWs includes an OLAP server layer where the
cubes are created and an OLAP front-end layer for cubes manipulation and
analysis. The same architecture can be used for SDWs counting on spatial
extensions for OLAP server and client.
The usual practice in implementing OLAP projects based on free
software relies on a well-known Mondrian OLAP server developed by
Pentaho (Pentaho 2013b). Mondrian is written in Java and requires an
XML fi le to defi ne a schema consisting of a cube with its dimensions
and measures. This cube can be queried later on by applying the Multi-
Dimensional Expressions (MDX) language. Schema and queries can
be written manually or the Mondrian Workbench interface with visual
components can be used instead.
Furthermore, a current stable version of Mondrian provides connections
to any JDBC data source and includes an interface specifi cation based on
XML/A, i.e., SOAP, as well as olap4j. In addition, Mondrian represents
memory-based architecture, i.e., reads data from the disk and copies it into
the cache. Although it can put limits on Mondrian's performance, it provides
an option of creating aggregated tables in the disk that can be used instead
of bringing the base data and performing the calculations in the cache.
Mondrian was extended with spatial features (Spatialytics 2013b)
allowing the defi nition of a vector geometry, as well as enriching MDX with
spatial capabilities. Its fi rst offi cial version called GeoMondrian was released
 
Search WWH ::




Custom Search