Database Reference
In-Depth Information
Time
Road Type
dateTime
...
Calendar
number
typeName
...
Day
Trajectory
Episode
date
week
isHoliday
...
number
startTime
endTime
startLocation
endLocation
...
route m( )
distance
duration
avgSpeed
Month
month
...
Car
number
model
year
fuel
...
Year
Car Type
year
...
name
...
Figure 4.3 An alternative partition of episodes with respect to road type.
The DW depicted in Figure
4.2
partitions a trajectory into
episodes
with
respect to days, roads, and districts. An alternative schema shown in Figure
4.3
partitions trajectories with respect to the road type in which they occur. For
example, a trajectory can be segmented into episodes occurring in highways,
national roads and regional roads. This partitioning is close to the notion of
episodes discussed in Chapter 1. Notice also that the time granularity in Fig-
ures
4.2
and
4.3
differs. In the former case, the granularity is day, although we
keep the movement track in the
route
measure with a timestamp granularity.
In the latter case, we relate each episode with its initial and final timestamps.
The choice among the two alternative data warehouse schemas depends on
application requirements and the typical OLAP queries to be addressed.
When trajectories are used as measures, the problem of aggregation arises.
In the examples of Figures
4.2
and
4.3
, we segmented the trajectories into epi-
sodes and kept their movement track in a geometry of type time-dependent
point. Thus, we can aggregate such episodes (or the whole trajectories) along
the different dimensions. An alternative approach for trajectory aggregation
aims at identifying “similar” trajectories and merging them in a class. This
aggregation may come together with an aggregate function, which may be the
count
function in the simplest case, although more complex ones may be
used. The main problem consists in adopting an appropriate notion of
trajectory
similarity
, through the definition of a similarity measure, for example, a
distance