Graphics Programs Reference
In-Depth Information
Save the image as a PDF and then open it in Illustrator. Follow the same
process that you've used previously to edit the step chart to your liking.
Design-wise, I got rid of the vertical axis altogether and directly labeled
each jump (with the Type tool). I also made evenly spaced tick marks, but
only labeled the years when there were changes.
When you have a
small dataset, it
can sometimes be
useful to label
points directly
instead of using a
value axis and
grid. This places
more emphasis
on the data, and
because there
aren't that many
points, the labels
won't clutter and
confuse.
Finally, I used a gray background. This was a personal preference, but
the background helps highlight the graph, especially placed within text.
It provides more surrounding space while not being annoyingly flashy. To
make a background behind the graph and text, instead of on the top cover-
ing everything, you need to create a new layer in Illustrator. You can do this
in, what else, the Layers window. Click the button to create a new layer. By
default, the layer will be placed on top, but you want it on the bottom, so
click-and-drag the new layer below Layer 1, as shown in Figure 4-45.
You can rename your layers, which can
be especially useful when you start to
design more complex graphics. I renamed
the new layer “background.” Then draw a
rectangle using the Rectangle tool. Click
and draw to make the size you want, and
then change color via the Color window.
FIGurE 4-45 Layer window in
Illustrator
Smoothing and Estimation
When you have a lot of data, or the data you have is noisy, it can be hard to
spot trends and patterns. So to make it easier, you can estimate a trend
line. Figure 4-46 shows the basic idea.
Draw a line that goes through the most points as possible, and minimize
the summed distance from the points to the fitted line. The most straight-
forward route is to create a straight fitted line using the basic slope-
intercept equation you probably learned in high school.
y = mx + b
Slope is m and the intercept is b . What happens when your trend is not
linear? It doesn't make sense to it a straight line to data that shows winds
and curves. Instead use a statistical method created by William Cleveland
and Susan Devlin called LOeSS, or locally weighted scatterplot smoothing.
It enables you to it a curve to your data.
Search WWH ::




Custom Search