Closed friendly closed 9 years ago
2 Possible solutions:
cotabplot(tab, layout = c(4,4), panel = cotab_fourfold) cotabplot(tab, layout = c(2,8), panel = cotab_fourfold)
Of course, the labels use a lot of space.
Currently, the drawing of labels cannot be suppressed in fourfold(), but would that be an option anyway? How to identify the strata?
David
On 2015-04-19 19:57, Michael Friendly wrote:
|tab <- xtabs(visit ~ practitioner + place + gender + insurance + chronicf, data=nmes.long) fourfold(tab, mfcol=c(4,4))|
Neither solution works here, so I guess I may leave this as it is.
The cotabplot
idea could work, if cotabplot used vertical labels for the rows, as ggplot does with faceting. I'm not saying you should do this now, but it would be an improvement for cotabplot in the long term. (I don't want to add unessential tasks delaying VCDR.)
I wouldn't suppress the labels entirely, but another improvement would be to allow the label to be either varname : value (as now) or just value.
I added an argument varnames= to rev. 806 which suppresses the variable names if set to FALSE. (That was trivial).
On 2015-04-20 15:51, Michael Friendly wrote:
Neither solution works here, so I guess I may leave this as it is. The |cotabplot| idea could work, if cotabplot used vertical labels for the rows, as ggplot does with faceting. I'm not saying you should do this now, but it would be an improvement for cotabplot in the long term. (I don't want to add unessential tasks delaying VCDR.)
I wouldn't suppress the labels entirely, but another improvement would be to allow the label to be either varname : value (as now) or just value.
— Reply to this email directly or view it on GitHub https://github.com/friendly/VCDR/issues/38#issuecomment-94456501.
FH-Prof. Priv.-Doz. Dr. David Meyer Institut für Wirtschaftsinformatik
Fachhochschule Technikum Wien Höchstädtplatz 5, 1200 Wien T: +43 1 333 40 77-394 F: +43 1 333 40 77-99 394 E: david.meyer@technikum-wien.at I: www.technikum-wien.at
Fig 11.42 uses mfcol() for a complex collection of fourfold plots, but the panel titles have the combinations of 3 variables represented, and overlap considerably. Is it possible to do this so that either the variable names are abbreviated or eliminated in the titles?