Exploring Data Visually - Data Points: Visualization That Means Something - page 154

VISUALIZING TIME SERIES DATA
Time passes. Things change, people change, and places change. You can feel
time through the sunrise and sunset, your clocks and watches, and the coffee
you need to drink when you wake up. When you visualize time series data, as
shown in Figure 4-15, your goal is to see what has passed, what is different, and
what is the same, and by how much. Compared to last year, is there more or
less? What are possible explanations for the increase, decrease, or nonchange?
Is there a recurring pattern, and is that good or bad? Expected or unexpected?
As with categorical data, the bar chart is a straightforward way to look at
data over time, except instead of categories on one of the axes, you use time.
Figure 4-16 shows the unemployment rate in the United States from 1948 to
2012, according to the Bureau of Labor Statistics. On top is the rate month-to-
month, and because there is a high point density, it looks like a continuous
area. On the other hand, the graph on the bottom shows only the unemploy-
ment rate in January of each year, which allows for space in between bars and
makes it easier to distinguish individual points.
In Chapter 1, you saw how car crashes vary over time, and how you can explore
time series data at different granularities. The same applies here. You can look
at data hourly, daily, annually, by decade, by century, and so on. Sometimes
the data format dictates the level of detail because the metric was measured,
say, only every 5 years. However, if for example you had measurements by
the hour, a high variability might obscure a trend that's more obvious if you
take a step back and look at your data by the day.
Usually the magnitude of change between segments of time is more interesting
than the value at each point. Although you can interpret trends from a bar
graph, you must visually calculate rates. You look at one bar and compare it
to the ones before and after.