1.6 Structure of the book
There is an old adage: “In theory it works in practice, in practice it doesn’t”. It might also be said that there is theoretical clarity and practical ambiguity. The applications in case studies come first in the book and the theory in the form of general advice arising from the case studies follows. The case studies can be read in any order, depending on what topics and graphics attract and interest. The second part pulls together ideas from the first part and offers an overview of the main factors affecting understanding graphics. If only the advice is of interest, just stick to the second part. (Ideally that will send readers back to the examples in the first part.) There are also brief summaries at the end of each chapter, listing some of the points made. The graphics are important and need as much attention or more as the text. Switching between studying graphics and reading text disrupts a smooth reading flow, but is worth the effort. This is a book that can be flicked through and dipped into.
The terms graphic, plot, and display all refer to graphic displays of data. Figure, chart, and graph also arise. Sometimes one seems to fit better than another, and there is no intention to suggest different meanings. The vocabulary of graphics is like the theory of graphics, flexible rather than rigid. This is also why much of the book’s advice is expressed in cautionary fashion. Words like ‘may’, ‘might’, ‘could’, ‘possibly’, ‘sometimes’ qualify what is said. Readers who are looking for clear-cut, straightforward rules may (!) be surprised. Exploring graphics in practice should convince them of the value of an open approach.
The layout of individual pages of the book has been strongly influenced by the need to ensure that the discussion of a graphic is always on the same or facing page as the graphic itself. This required minor adjustments to the size of graphics, to the text, and to the order of text and graphics. Some sections have been expanded a little and some cut back a bit. A few pages are shorter than others, where it was thought better to leave space.
1.6.1 Part I Graphics in Action
Graphics are an integral part of all of Data Analysis: they are not just for presenting results. Looking at individual graphics on their own is like testing your eyesight by identifying individual letters or testing your hearing by recognising individual words. It is valuable for assessing your knowledge of the graphics, but not the same as looking at them as a group in context. Your mind can work out what words you are seeing without identifying each letter and can work out the meaning of what is being said to you without hearing every word clearly. The same holds for groups of related graphics. It is like having a conversation with the data.
Graphics are part of data exploration and of the quality control of data. Reporting a case study ignoring these important activities would be misleading, while reporting false trails and repeated analyses would be confusing and irritating. So the reporting of case studies here tries to follow a middle path. If the order seems surprising, such as only correcting some bad data late on, this is because sometimes this is the way analyses proceed. With large datasets new data problems may continually emerge. Not every variable is checked initially, they are investigated as they are needed. New details become apparent, sometimes relevant, sometimes not. Details can enlighten and enliven, shedding new light and awakening interest. Documenting case studies as if they consisted of steady progress from beginning to end, culminating in impressive conclusions, would be unrealistic. Case studies are not like mathematical proofs. Most of the case studies suggest further work and not all have as substantial conclusions as might have been hoped for initially.
1.6.2 Part II Advice on Graphics
The chapters in this part cover general topics on graphics taking examples from the case studies to demonstrate the ideas. The illustrations are sometimes smaller versions of the graphics used in the case studies to make a particular point. They are cross-referenced so that it is easy to check the original versions. There are chapters on provenance and quality (investigating the sources of data and what has been done with the data); wrangling (cleaning, organising, and restructuring data); colour; setting the scene (layout and formatting); ordering, sorting, and arranging; interpretation (paying attention, understanding, checking); varieties of plots and developing plots. This part could be read first to see what factors the book regards as important and to get a summary of the book’s advice, but readers would miss out on getting the experience of graphics in action.