Visualizing Proportions - Data Points: Visualization That Means Something - page 178

Wrapping up
The main thing that sets proportions apart from other data types is that
they represent parts of a whole. each individual value means something,
but so do the sum of all the parts or just a subset of the parts. The visual-
ization you design should represent these ideas.
Only have a few values? The pie chart might be your best bet. Use donut
charts with care. If you have several values and several categories, con-
sider the stacked bar chart instead of multiple pie charts. If you're looking
for patterns over time, look to your friend the stacked area chart or go for
the classic time series. With these steady foundations, your proportions
will be good to go.
When it comes time to design and implement, ask yourself what you want
to know about your data, and then go from there. Does a static graphic tell
your story completely? A lot of the time the answer will be yes, and that's
fine. If, however, you decide you need to go with an interactive graphic,
map out on paper what should happen when you click objects and what
shouldn't. It gets complicated quickly if you add too much functionality, so
do your best to keep it simple. Have other people try interacting with your
designs to see if they understand what's going on.
Finally, while you're programming—especially if you're new to code—
you're undoubtedly going to reach a point where you're not sure what to
do next. This happens to me all the time. When you get stuck, there's no
better place than the web to find your solution. Look at documentation if
it's available or study examples that are similar to what you're trying to do.
Don't just look at the syntax. Learn the logic because that's what's going to
help you the most. Luckily there are libraries such as Protovis and Flare
that have many examples and great documentation.
In the next chapter, we move towards deeper analysis and data interpreta-
tion and come back to your good statistical friend. You put R to good use as
you study relationships between data sets and variables. Ready? Let's go.