Chapter 14. Topics in graphical analysis
This chapter covers
- Techniques for time-series plots
- Techniques for multivariate data sets
- Techniques to improve visual perception
The previous chapter discussed some fundamental graphical methods that are very generally useful. In this chapter, I’ll select a few tasks that are more specific and explore them in greater depth.
Because of their immense practical importance, our first topic will be the handling of time series. In particular, you’ll see some clever ways to smooth a time series using only gnuplot commands. The next topic concerns multivariate data sets: data sets with more than two or three different attributes. A variety of techniques are available for such data, and some of gnuplot’s relatively new features facilitate the visual exploration of such data.
Next, we’ll discuss ways to enhance the visual perception of graphs. You can aid analysis and discovery by preparing graphs such that noteworthy features are more easily recognized. This topic isn’t frequently discussed, but some awareness of the issues involved can greatly improve your results.
Time-series plots are probably the most common graph in practice: whenever you want to monitor a quantity, you’re dealing with a time-series plot of some form. This section introduces some operations that are frequently useful when dealing with time series and shows you how to coax gnuplot into performing those operations.