HMSC UseR Meeting, 13 November 2015
Protip # 1: simple is better
Avoid the "Christmas Letter Syndrome"
- Maintain uniformity in font size, style, and color
- Avoid text in italics or bold
Protip # 2: adopt scientific graphics principles
Edward Tufte's (2001) Rules
- Maximize the data-to-ink ratio
- Erase non-data ink
- Increase the data density
- Label the figures, avoid using legends
(Shout-out to @EdwardTufte and @TrevorABranch)
Protip # 2 (cont.): adopt scientific graphics principles
- Stephen Few's practical advice on data visualization
- William Cleveland's "Elements of Graphing Data" (1994)
- Leland Wilkinson's "Grammar of Graphics" (2005), implemented in ggplot2
- Add-on library ggthemes implements additional "popular" themes for ggplot2: Solarized, The Economist, Tableau, Stata, Excel, and Few's
- Winston Chang's R Graphics Cookbook
Protip # 3: choose color schemes judiciously
- Stephen Few’s Practical Rules for Using Color in Charts
- Use color-blind-friendly palettes
- Use ColorBrewer palettes, implemented in RColorBrewer
- Use spectral palettes with good perceptual characteristics:
- RColorBrewer (sequential, divergent)
- Viridis
- Avoid rainbow-like palettes (Matlab's "jet" replaced w/ "Parula")
Protip # 4: avoid problematic graph types
- Pie charts (use dot plots)
- Bar charts (use dot plots)
- Frequency histograms (use density plots)
- 3-D scatterplots (use 2-D scatterplots, "condition" [group] the data by a third variable and visualize with trellis in lattice or faceting in ggplot2)
Protip # 5: strive for small graphics files
- For drafts, PPTs, web: export PNG version (100 KB - 1 MB)
- For publisher: TIFF with LZW compression (1-5 MB) for raster or EPS for vector
Protip # 6: do not export as PDF
- PDF is not a graphics file format!
- Graphics with large amounts of data cause significant slowness in PDF display
- However, exporting as PDF offers significant conveniences, especially at the data exploration/draft stage. A set of plots can be exported to a single file with multiple pages. PDF files are easily viewable and offer high-quality rendering for vector graphics.(Shout-out to @TrevorABranch)
References & Resources
Cleveland, WS (1994) The Elements of Graphing Data, 2nd Ed. Hobart Press, Summit, New Jersey.
Kappel, E (2009) Common graphics misconceptions. Oceanography 22: 7–9. doi:10.5670/oceanog.2009.53
Tufte, ER (2001) The Visual Display of Quantitative Information, 2nd Ed. Graphics Press, Cheshire, Connecticut.
Weissgerber TL, Milic NM, Winham SJ, Garovic VD (2015) Beyond bar and line graphs: time for a new data presentation paradigm. PLoS Biol 13(4): e1002128. doi:10.1371/journal.pbio.1002128
Wilkinson, L (2005) The Grammar of Graphics, 2nd Ed. Springer-Verlag, New York.