Graphics Programs Reference
In-Depth Information
However, where the bars look like a continuous unit (refer to Figure 4-16),
it's easier to distinguish changes because you can see the slope, or the rate
of change in between points. It's even easier to see the slope when you use
a continuous line, as shown in Figure 4-17. The line chart shows the same
data as the bar graph on the same scale, but change is directly displayed via
direction as a visual cue.
A dot plot can be used in the same way, as shown in Figure 4-18. Again, the
data and axes are the same and the visual cue is different.
Like bar charts, dots put focus on each value, and trends can be harder to see.
Although in this example, there are enough data points, so you don't need to
mentally fill in the gaps. If the data were more sparse, such as in Figure 4-19,
changes are less obvious.
When you connect the sparse dots with a line, as shown in Figure 4-20, the
focus of the plot shifts again.
If you care more about an overall trend than you do about the more spe-
cific monthly variability, you can fit a LOESS curve to the dots, as shown in
Figure 4-21, instead of connecting every dot. The closer you fit the curve to
the dots, the more it resembles Figure 4-17.
Note: LOESS (or LOWESS ) stands for locally
weighted scatterplot smoothing . It's a statisti-
cal technique created by William Cleveland, which
fits a polynomial function to a subset of the data
at different points. When combined, they form a
continuous line.
Of course, the chart style you choose depends on your
data, and although it might seem like a grab bag of options
at first, you get a feel for what type of chart to use with
practice. It's not an exact science (or computers could do
all the work) and options can vary a lot even if you have
datasets that look similar.
For example, the previous charts on unemployment rate provide a historical
view of the past few decades. You can see peaks and valleys, periods of reces-
sion such as in 2001 and from 2007 to 2009, and an overall picture of changing
rates. If you were only interested in the five highest peaks and what happened
immediately after them, the exploration would take a different route.
Search WWH ::




Custom Search