Global Positioning System Reference
In-Depth Information
data related to zones of fl ooding, fi re, or earthquake risks, localization of
fi re stations, clinics, schools, crop, and vegetation distributions, among
others? Could the development of SOLAP solutions with these kinds of
data extended with other conventional data help the users at the different
levels of administration to improve the planning of local development
or risk prevention? Could professional from other fi elds, e.g., agronomy,
biology, medicine, hydrology, benefi t from SOLAP solutions for exploring
spatial and conventional data in a multidimensional manner before other
kinds of analysis are applied?
Furthermore, there are other aspects that make the acceptance of
conceptual modeling diffi cult, before the SDW and SOLAP implementations
take place: (1) lack of tradition in GIS and DW communities to use conceptual
models, relying instead on either shape fi les (for GIS) or star or snowfl ake
logical schemas (for DWs), (2) a variety of research proposals with their
own specifi cation and presentation, (3) lack of a well-known and accepted
model for representing DW or SDW structures, and (4) the proposed models
seldom include rules that are required for transformations from conceptual
to logical or physical levels in order to implement a SDW.
Regarding implementation, there is a clear tendency of converging
GIS and DBMS technologies by having a spatial extension in DBMSs
and promoting geo-databases (that in fact are spatial DBMSs) in the GIS
community (ESRI 2012). As indicated for conventional DW (Costa et al.
2006; Golfarelli and Rizzi 2009; Jensen et al. 2010), through an adequate
physical design (e.g., aggregated tables, indexes) query performance
may be improved. This aspect should also be investigated in the context
of SDWs. Furthermore, the implementation process should consider
whether normalized (snowfl ake) or denormalized (star) tables are a better
choice. For conventional data, there are several recommendations (e.g.,
Jensen et al. 2010; Martyn 2004) and, fortunately, Costa et al. (2010), based
on several experiments with spatial dimensions represented as star and
snowfl ake schemas, also give advice that may be useful before choosing
either option. In addition, star and snowfl ake schemas for conventional
DW were extended (Jensen et al. 2010), thus enabling the implementation
of other features, e.g., many-to-many relationships between facts and
dimensions, degenerate dimensions, and different kinds of hierarchies, e.g.,
parent-child, non-covering, non-strict. Whether these features are required,
useful, and can be implemented for SDWs, it is still an open research topic.
Nevertheless, based on our experience, there are still several aspects that
may infl uence negatively the possibility to consider spatial DBMSs in the
SDW implementation:
• Spatial DBMSs are relatively a new technology, thus, they are not well
known outside the small circle of implementers. There are still not
Search WWH ::
Custom Search