Graphics Reference
In-Depth Information
trend even than the general pattern of improving performance as you move
southward, which you might imagine could be attributed to weather.
Another trend that seems apparent is that larger airports seem a bit more
likely to be delayed than smaller airports. This is particularly evident on
the West Coast and the area around Florida. One explanation might be
that larger volumes of air traffic tend to be more susceptible to congestion.
However, given that this phenomenon seems fairly exclusive to areas where
flights are often on time, a more likely reason might be that the smaller
airports in these areas do not have as many flight connections to airports
that tend to incur delays.
Track Aggregation
So far, you have analyzed flight delays as a consumer. You may have
concluded that you are best off by booking connections along coastal routes
and avoiding particular airports. Taking it further, you might gather
additional data to analyze time of day and seasonal differences, and
correlate with weather patterns, or compare patterns between particular
airlines or aircraft types. Given the observations made, you could develop
algorithms to forecast the best flight options when booking and use
visualization to help explain why certain routes are better than others.
However, as a consumer, you only need to optimize your choice of flight
segments, similar to how you might choose roads to take to work, but
with fewer options. From the perspective of an airline or airport air traffic
controllers, routing decisions are decidedly more complex. An optimal path
through airspace can, in theory, be different for each flight, each day.
When route data is not bound to path segments but rather is tracking freely
across a spatial field, it is often more feasible to aggregate the data in the
form of a two-dimensional (2D) or three-dimensional (3D) field. Aircraft
and ship tracks are good examples of this kind of data.
Figure 13-11
shows a visualization of ship tracks by frequency over a little
more than a year. This aggregation technique lends itself well to a tiled,
multi-resolution visualization approach with zooming and panning, similar
to what you would expect of an online map.
