spiderBoxplot {GCDkit} | R Documentation |
Normalization of geochemical data by the given standard (optionally also one of the samples) and spiderplot plotting. No individual patterns are drawn; instead, the statistical distribution of each element is portrayed by a boxplot.
spiderBoxplot(norm = NULL, which = rep(TRUE,nrow(WR)), doublenorm = FALSE, norm2 = "", ymin = NULL, ymax = NULL, bpplot = TRUE, col = "lightgray", log = TRUE)
norm |
a character string specifying the model. |
which |
specification of the samples to be plotted. |
doublenorm |
logical; should be the normalization employed? See details. |
norm2 |
name of the variable for the second normalization. |
ymin, ymax |
y range of the diagram. |
bpplot |
logical; if FALSE, boxplot box (instead of box and percentile plot) is shown. |
col |
fill colour. |
log |
logical; should be the y axis scaled logarithmically? |
The parameter 'norm' is an optional search pattern to
query the available normalizing model names. It can contain a
substring or even a regular expression. The function fails
if no matches are found or the search is ambiguous. See selectNorm
for details.
The samples to be plotted can be selected based on combination of three
searching mechanisms (by sample name/label, range or a Boolean condition) - see
selectSamples
for details.
For choosing the correct normalization values serves the auxiliary function
selectNorm
, which is the same as in ordinary spiderplots.
If the user desires so, the data can be normalized by a sample
present in the dataset. Then the elements to be plotted and their order is to
be specified, as well.
Optionally, double normalization can be used. Trace-element data are first
normalized by the given standard, then by the normalized content of the selected
element in each analysis to eliminate effects of fractional crystallization
(Thompson et al. 2003, Pearce et al. 2005, Pearce and Stern 2006).
See spider2norm
for details.
Distributions of individual normalized elements are plotted in the form of boxplot or box and percentile plot (Esty and Banfield 2003).
In both cases the box denotes 50% of the population (both quartiles),
the horizontal line in the middle is a median and the whiskers denote the overall
range. For boxplot this is without outliers.
See manual entry for 'boxplot
' and 'bpplot.my
' for further details.
Printed are number of observations, missing values, mean, standard deviation, minimum, 25% quartile, median (=50% quartile), 75% quartile and maximum.
results |
numeric matrix with statistical data for individual elements. |
This function IS NOT Figaro-compatible. It means that the set of diagrams cannot be further edited in GCDkit (e.g. tools in "Plot editing" menu are inactive).
Vojtěch Janoušek, vojtech.janousek@geology.cz
Esty, WW & Banfield JD (2003) The Box-Percentile Plot. Journal of Statistical Software 8 (17)
Pearce JA, Stern RJ (2006) Origin of back-arc basin magmas: Trace element and isotope perspectives. Back-Arc Spreading Systems: Geological, Biological, Chemical, and Physical Interactions. Geophysical Monograph Series 166. American Geophysical Union, pp 63-86
Pearce JA, Stern RJ, Bloomer SH, Fryer P (2005) Geochemical mapping of the Mariana arc-basin system: implications for the nature and distribution of subduction components. Geochem Geophys Geosyst 6: doi: 10.1029/2004GC000895
Thompson RN, Morrison MA, Dickin AP, Hendry GL (1983) Continental flood basalts... Arachnids rule OK? In: Hawkesworth CJ, Norry MJ (eds) Continental Basalts and Mantle Xenoliths. Shiva, Nantwich, pp 158-185
For the syntax of the setup file with normalizing values and adding new
normalization schemes
see selectNorm
; for further applications of
'spider
' see spiderByGroupPatterns
,
spider2norm
and spiderByGroupFields
.
data(atacazo) accessVar("atacazo") spider(WR,"Boynton",ymin=1,ymax=100) spiderBoxplot("Boynton",col="yellow",bpplot=FALSE) spiderBoxplot("Boynton",col="yellow",bpplot=TRUE) spiderBoxplot("Primordial Wood",doublenorm=TRUE,norm2="Y", col="khaki",ymin=0.05,ymax=1000,bpplot=TRUE)