During my professional career I have never been told how to make good graphics although extracting essential information from datasets is an advanced (and necessary skill). Fortunately, however, there are some excellent books that cover graphical display. The first one I came across was Michael J Campbell and David Machins’ “Medical Statistics – A commonsense approach” (see pages 44ff and 58 for “increasing data ink”) that gives a lot of useful advices how to improve figures. The next book that I found influential was Bill Cleveland’s 1993 book “Visualizing Data” who introduced into R and S multidimensional lattice graphics (also covered 2005 in “R Graphics” of Paul Murrell published by Chapman & Hall/CRC). I used this technique extensively in my 2005 PLoS paper on the worldwide distribution of allergy. At the moment I am reading the new book “Graphics of large datasets” by Antony Unwin, Martin Theus, Heike Hofmann which seems to be finally the masterclass of displaying data. Yea, yea.
An examples how to improve a barchart (yes – I resisted to start with a 3D barchart but the cluttering colors and grids are hopefully good to see).
A new Nature Nascent entry: “The way we present genomic and proteomic data on the web sucks”