Database Reference
In-Depth Information
Time
District
Zone
date
week
isHoliday
...
name
population
area
...
number
name
...
Road
Calendar
number
name
...
Episode
Month
month
...
Trajectory
route m( )
distance
duration
avgSpeed
number
distance
duration
startTime
endTime
startLocation
endLocation
...
Quarter
quarter
...
Car
number
model
year
fuel
...
Year
Car Type
name
...
year
...
Figure 4.2 An example of a trajectory data warehouse.
4.3 Running Example
We introduce next the running example that will be used throughout this chapter.
The Italian city of Milano has one of the highest rates of car ownership in Europe.
Since this induces many problems, a DW can be useful for understanding and
analyzing traffic data so that corrective measures may be taken. Spatial data in
the warehouse include the road network, the political division of the city into
zones and districts (administratively, the city is divided into nine zones, each zone
encompassing a number of districts), and the trajectories themselves. Nonspatial
data include the characteristics of the car performing the trajectory. Figure
4.2
shows the
conceptual
schema depicting the above scenario using the MultiDim
model due to Malinowski and Zimanyi (although any other conceptual model
could be used instead). Note that to support spatio-temporal data, we extended
the MultiDim model with
time-dependent
(or
moving
)
types
, which capture
the evolution over time of base types (e.g., real, integer) and spatial types.
For details about these data types and their operators, we refer the reader to
Chapter
3
.
When building a data warehouse, the data to be analyzed (in our case tra-
jectories) determine the
facts
and associated
measures
. An important question
then is to determine the axes of analysis, or
dimensions
, that will be used for
