Geography Reference
In-Depth Information
must then be employed. In this section, the flow map technique was embedded
into another spatial exploratory data analysis technique called the comap.
The comap (Brunsdon 2001) and its older, more generic cousin, the coplot
(Cleveland 1994), provide an effective means of visualising multivariate data
in order to assist with uncovering previously hidden patterns embedded within
complex data. The basic idea is that a bivariate subset of raw data is selected
based upon some condition (for example, a defined range of rainfall or a
categorical variable such as weekend or weekday). This data is then plotted
either in raw form in a scatter plot panel or a mapped kernel density surface.
The comap and coplot embrace the small multiple principle (Tufte 1991) by
using multiple panes of similar and typically overlapping regions, in order to
illustrate how gradual changes can be observed as a function of external
variables.
Embedding the flow map into the comap to form a combined technique
that represents a straightforward extension of the two techniques which allows
multivariate exploration of origin-destination data which otherwise would not
be readily possible.
3.3. CityCycle Flow Dynamics
We examine the spatial flows of City Cycle over a 24-hour period and the
effect of specific calendar events (i.e. weekdays, weekends, public and school
holidays) on the trip patterns. The flow maps of CityCycle between Brisbane's
inner suburbs are displayed by Figure 3. The number of trips between suburbs
represented using a variable line thickness, where the width of the line is
proportional to the total flow volumes. Origin-destination pairs generating less
than 200 trips in total were not mapped in order to preserve graphical clarity
and highlight the main suburb-to-suburb interactions. The choropleth
classification of suburbs represents the within-suburb flows (i.e. where the
bicycle release and return are in the same suburb), with the darker colour
representing the higher number of internal trips.
As Figure 3 shows, a high proportion of trips taking place over relatively
short distances, along with a high degree of interaction between adjacent
suburbs. The number of trips appears to be higher (but not concentrated)
between the Brisbane Central Business District (CBD) and the immediate
surrounding suburbs, whilst the trips between suburban locations are lower. In
regards to self-containment, a number of suburbs exhibit relatively high levels
of within suburb trips.
