Global Positioning System Reference
In-Depth Information
are represented: the star schema uses a fl at table (i.e., all attributes forming
a hierarchy are included in the same table, e.g.,
Product name, SubCategory
Name, Category Name
in the
Product
table in Fig. 1), while the snowfl ake
schema relies on representing a hierarchy as a normalized structure (e.g.,
tables
Store
,
District
,
County
, and
Province
in Fig. 1).
Hierarchies are important in analytical applications since they
represent data at different levels of granularity (e.g., district, country, and
province) allowing analyses at different levels of detail and exploiting
OLAP systems to their full capabilities. When a hierarchy is traversed
from fi ner to coarser levels, measures are automatically aggregated, e.g.,
moving in a hierarchy from a
Store
to a
District
level will give aggregated
sales measures for different districts based on sales of their corresponding
stores. This operation is called
roll-up
, while another opposite operation that
provides more detailed data from aggregated values is called
drill-down
.
The usual practice during the aggregation is to apply the sum operators.
However, some measures, called
non-additive
, cannot use this operator,
since the aggregation is meaningless, e.g., the
Price
measure in Fig. 1; other
measures, called
semi-additive
, e.g.,
Quantity
in the
Inventory
DW, can be
aggregated in any dimension, but the
Time
dimensions, i.e., it is incorrect
to add the quantity of existing products in inventory considering different
periods of time.
As a consequence of the growing demand to incorporate spatial data
into the decision-making process, SDW and SOLAP models based on
multidimensional view of data have emerged (e.g., Bédard et al. 2009;
Bimonte et al. 2010; Silva et al. 2010; Damiani and Spaccapietra 2006;
Gómez et al. 2008; Jensen et al. 2004; Malinowski and Zimányi 2008;
Pourabbas 2003). SDW and SOLAP strongly emphasize on spatial data
represented in dimensions and measures, as well as on the necessity to
include different spatial functions including the ones required for spatial
measure aggregations.
Case study: Analysis of Cancer Incidence and Demographics
Through the years, the Central American Population Center (CCP in
Spanish) at the University of Costa Rica has collected data about population,
death and birth rates, incidence of different types of cancers, read/write
literacy levels, professional occupation, and housing equipment, among
others (CCP 2013). This data is provided by various public institutions;
for example, the National Institute of Statistics and Census publishes
census data from different time periods or several research centers at the
University of Costa Rica deliver necessary data to expand social awareness
about important issues that allow the country to improve its situation. Since
available data uses different formats, CCP stores it in different systems.
Search WWH ::
Custom Search