Visualizing the meaning of log-scales

Plotting in log-scale is often the only way of displaying a given dataset. Unfortunately, log-scales are more difficult to understand even for trained eyes. Often, it is used the formula "note the log-scale of y-axis" in the figure legend. However, it hardly makes the plot easier to follow for log-adverse people. This plot shows an elegant way to display the untransformed data and, at the same time, benefiting from the illustrative power of a log-scale. 

Reference Hebert, P.D.N., Ratnasingham, S. and deWaard, J.R. (2003) Barcoding animal life: cytochrome c oxidase subunit 1 divergences among closely related species. 2003 Proc. R. Soc. Lond. B S96-S99 270 doi: 10.1098/rsbl.2003.0025

Read more...

Breaking plotting rules

The interesting thing about this plot is the reversed x-axis: from left to right "vegetation height" goes from tall to short vegetation. Note the right-hand arrow indicating this unusual x-axis scale. The authors explain that this was done to "...emphasize that islands with low vegetation are more dissimilar from Staniel than are islands with high vegetation.".

More generally, this plot illustrates that there are apparently fixed plotting rules that can be (carefully) broken in a very interesting way.

Reference J.B. Losos, K.I. Warheit and T.W. Schoener (1997) Adaptive differentiation following experimental island colonization in Anolis lizards. Nature 387: 70-73.  (download pdf)

Read more...

Video-graph

A fantastic way to visualize spatial patterns in time. This can not be done in the printed paper, but it can be introduced as a .gif file in the supplementary material.

Figure S1. Movie of historical defoliation caused by the Larch Budmoth, Zeiraphera diniana in the European Alps, 1961-1998. Inset graph shows the proportion of all areas that were defoliated in each year. Defoliation polygons were aggregated to form raster estimates of the proportion of 20 x 20 km grid cells that were defoliated. Bubbles are proportional to the amount of defoliation in each cell in each year. Visual inspection of the movie suggests the existence of repeated outbreak waves moving from west to east.

Reference Bjørnstad, O.N., Peltonen, M., Liebhold, A.M. & Baltensweiler, W. (2002) Waves of larch budmoth outbreaks in the European alps. Science 298: 1020-1023. (download pdf) (Supplement) (paper in Science website).
Related info: Ottar Nordal Bjørnstad webpage
Input idea courtesy of Pablo Almaraz

Read more...

Shaded confidence intervals

Confidence intervals are usually displayed as dashed lines around the main line. This often creates a messy graph, particularly when plotting more than one trend. The use of shaded areas shown here is an elegant alternative. This allows combining different information in the same graph (e.g. comparing model predictions with empirical data), and may be specially interesting when small (e.g. inset) panels are required. Note that an extension of this technique would be to use different gray intensities to display different confidence intervals around the same main trend.




Figure legend. Black lines correspond to field data, while the white line and the shaded area represent the median and the 90% credible intervals obtained from samples of the posterior distribution of selected models. 

Figure from a previous version of the ms. Martínez, I., Wiegand, T., Camarero, J.J., Batllori, E. & Gutiérrez, E. (2011) Disentangling the formation of contrasting tree-line physiognomies combining model selection and Bayesian parameterization for simulation models. American Naturalist 177: E136-E152. (link)

Input courtesy of Dr. Isabel Martínez Cano

Read more...

Table as inset

A powerful combination when space is a big constraint or when information on a table and a graph needs to be interpreted together.

Reference Diabaté et al. (2009) Spatial swarm segregation and reproductive isolation between the molecular forms of Anopheles gambiae. Proc. R. Soc. London B Published online before print September 4, 2009, doi: 10.1098/rspb.2009.1167. (download free pdf)

Read more...