sinhrks / ggfortify Goto Github PK

For some reason when I try to plot labels I can't ever the actual labels even after renaming the rows

There are some classes which doesn't hold original data itself, but stores a call signature when the model was created. If autoplot can retrieve original data from parent environment, we can omit the explicit data keyword.

Some standard function behave like above, but it can be an unexpected result if the original variable has been changed. Thus this behavior should be explicitly turned on by global option (or argument).

ggfortify gives different PCA loadings in biplot when compared with stats::biplot ouput. This is also reported here.

Modify .gitignore to reduce package file size.

From yesterday, following error occurs on Travis CI repeatedly. Maybe any dependencies has been changed? Strangely I can't reproduce the issue on my local.

The branch is unmodified fork of current master:

• https://travis-ci.org/sinhrks/ggfortify/jobs/88574863

* creating vignettes ... ERROR
Loading required package: proto
Loading required package: ggplot2
Warning in rgl.init(initValue, onlyNULL) :
  RGL: unable to open X11 display
Warning in fun(libname, pkgname) : Error in 'rgl_init'
Don't know how to automatically pick scale for object of type complex. Defaulting to continuous
Don't know how to automatically pick scale for object of type complex. Defaulting to continuous
Quitting from lines 112-121 (plot_pca.Rmd) 
Error: processing vignette 'plot_pca.Rmd' failed with diagnostics:
Discrete value supplied to continuous scale
Execution halted

The error message says it is related to lfda block.

library(lfda)

# Local Fisher Discriminant Analysis (LFDA)
model <- lfda(iris[-5], iris[, 5], 4, metric="plain")
autoplot(model, data = iris, frame = TRUE, frame.colour = 'Species')

# Kernel Local Fisher Discriminant Analysis (KLFDA)
model <- klfda(kmatrixGauss(iris[-5]), iris[, 5], 4, metric="plain")
autoplot(model, data = iris, frame = TRUE, frame.colour = 'Species')

Hi
How can I change the title of the legend? Right now it is always called strata

I tried scale_color_discrete(name = xlab) but then I got 2 legend:

Hi

I'm trying to plot acf outputs with ggplot2, and trying to use ggfortify, but get the error below.

Best,
Richard

library(ggplot2)
library(ggfortify)
Loading required package: scales
Error in unloadNamespace(package) :
namespace ‘scales’ is imported by ‘ggplot2’ so cannot be unloaded
Error in library(pkg, character.only = TRUE, logical.return = TRUE, lib.loc = lib.loc) :
Package ‘scales’ version 0.2.4 cannot be unloaded

Using R on Ubuntu Trusty: R: 3.2.0-4trusty0

As Hadley tweeted, forecast package will natively support ggplot2. That's a great news per se, but this means that they implemented their own fortify and autoplot methods, which conflict with ones in ggfortify package.

dplyr::intersect(ls(asNamespace('ggfortify')), ls(asNamespace('forecast')))
#> [1] "autoplot.acf"           "autoplot.Arima"         "autoplot.decomposed.ts" "autoplot.ets"           "autoplot.forecast"     
#> [6] "autoplot.stl"           "autoplot.ts"            "fortify.forecast"       "fortify.ts"  

I tried the dev version of forecast package and found the error (Sorry, I haven't dig into the detail yet... Please forgive me to just drop the console log here.):

library(ggplot2)
library(forecast)

# no problem here
ggplot(AirPassengers, aes(x, y)) + geom_forecast()

library(ggfortify)
#> Loading required package: proto

# fail!
ggplot(AirPassengers, aes(x, y)) + geom_forecast()
#> Error in eval(expr, envir, enclos) : object 'x' not found

Session info:

devtools::session_info()
#> Session info --------------------------------------------------------------------------------------------------------------------------
#>  setting  value                       
#>  version  R version 3.2.3 (2015-12-10)
#>  system   x86_64, linux-gnu           
#>  ui       RStudio (0.99.878)          
#>  language (EN)                        
#>  collate  en_US.UTF-8                 
#>  tz       <NA>                        
#>  date     2016-02-18                  
#> 
#> Packages ------------------------------------------------------------------------------------------------------------------------------
#>  package    * version    date       source                               
#>  assertthat   0.1        2013-12-06 CRAN (R 3.2.2)                       
#>  colorspace   1.2-6      2015-03-11 CRAN (R 3.2.2)                       
#>  DBI          0.3.1      2014-09-24 CRAN (R 3.2.2)                       
#>  devtools     1.10.0     2016-01-23 CRAN (R 3.2.3)                       
#>  digest       0.6.9      2016-01-08 CRAN (R 3.2.3)                       
#>  dplyr        0.4.3.9000 2015-10-01 Github (hadley/dplyr@dd10ccd)        
#>  forecast   * 7.0        2016-02-18 Github (robjhyndman/forecast@451cb34)
#>  fracdiff     1.4-2      2012-12-02 CRAN (R 3.2.2)                       
#>  ggfortify  * 0.1.0      2015-11-30 CRAN (R 3.2.3)                       
#>  ggplot2    * 2.0.0      2015-12-18 CRAN (R 3.2.2)                       
#>  gridExtra    2.0.0      2015-07-14 CRAN (R 3.2.2)                       
#>  gtable       0.1.2      2012-12-05 CRAN (R 3.0.0)                       
#>  labeling     0.3        2014-08-23 CRAN (R 3.2.2)                       
#>  lattice      0.20-33    2015-07-14 CRAN (R 3.2.1)                       
#>  magrittr     1.5        2014-11-22 CRAN (R 3.2.2)                       
#>  memoise      1.0.0      2016-01-29 CRAN (R 3.2.3)                       
#>  munsell      0.4.3      2016-02-13 CRAN (R 3.2.3)                       
#>  nnet         7.3-12     2016-02-02 CRAN (R 3.2.3)                       
#>  plyr         1.8.3      2015-06-12 CRAN (R 3.2.2)                       
#>  proto      * 0.3-10     2012-12-22 CRAN (R 3.0.0)                       
#>  quadprog     1.5-5      2013-04-17 CRAN (R 3.2.2)                       
#>  R6           2.1.2      2016-01-26 CRAN (R 3.2.3)                       
#>  Rcpp         0.12.3     2016-01-10 CRAN (R 3.2.3)                       
#>  scales       0.3.0.9000 2015-09-21 Github (hadley/scales@7fc502c)       
#>  tidyr        0.4.1      2016-02-05 CRAN (R 3.2.3)                       
#>  timeDate   * 3012.100   2015-01-23 CRAN (R 3.2.2)                       
#>  tseries      0.10-34    2015-02-20 CRAN (R 3.2.2)                       
#>  zoo        * 1.7-12     2015-03-16 CRAN (R 3.2.2)  

Hey,
Thanks for building this awesome package that helps working with ggplot2 in all circumstances ;)

There are quite a number of methods for sp objects in ggfortify but none for raster objects.
The rasterVis package include a fortify()-like function in its function gplot() .
I was thinking of implementing something similar in ggfortify.

• Once tag & release current master as v0.0.1
  • https://github.com/sinhrks/ggfortify/releases/tag/v0.0.1
• Revise doc descriptions and formats
• Change fortify and autoplot argnames to be compat with {ggplot2}
  • Confirm all the current examples work with warning after the change
  • Create test_compat and move previous examples.
  • Change all the examples to use new argnames
• Release pre0.0.2 and submit it to CRAN
  • ftp://cran.r-project.org/incoming/
• Fix indicated points and resubmit (#23)
• Fix indicated points and resubmit (#25)

Hi,

please, when i try to run

library(ggfortify)
library(survival)
fit <- survfit(Surv(time, status) ~ sex, data = lung)
autoplot(fit)

gave this error:
Error: GeomConfint was built with an incompatible version of ggproto.
Please reinstall the package that provides this extension.

thank you so much

The final line of fortify.survfit, post_fortify(d), seems to be a no-op.

> d2 = post_fortify(d)
> identical(d, d2)
    [1] TRUE

Should it just be removed? This might be the same as #102.

Hi,

Is it possible to add the PC1 and PC2 scores to the axes labels (like in ggbioplot: https://github.com/vqv/ggbiplot)?

Thanks

Details here.

@sinhrks Could you look into this when you get a chance?

There are few plotting functions using grieExtra::grid.arrange internally, such as autoplot.lm. These function doesn't have return value, thus we cannot decorate it later.

Consider a better way to allow decoration.

Using R v3.2.3 w/ ggfortify 0.1.0 and ggplot2 2.0.0 fails

> autoplot(survfit(Surv(time, status) ~ sex, data = lung))
Error in get(x, envir = this, inherits = inh)(this, ...) : 
  unused argument (TRUE)

Other conversation about this at: tidyverse/ggplot2#1458

• http://cran.r-project.org/web/packages/broom/index.html

Utilize {broom} as much if outputs data.frame which has enough information for plotting.

• aareg_tidiers
• anova_tidiers
• cch_tidiers
• confint_tidy
• coxph_tidiers
• cv.glmnet_tidiers
• felm_tidiers
• gam_tidiers
• glmnet_tidiers
• lme4_tidiers
• multcomp_tidiers
• nls_tidiers
• pyears_tidiers
• ridgelm_tidiers
• rlm_tidiers
• sexpfit_tidiers
• smooth.spline_tidiers
• sp_tidiers
• survfit_tidiers
• survreg_tidiers
• tidy.manova
• tidy.map
• tidy.TukeyHSD

Affects to backward compat

autoplot should not be affected. Column names output from fortify may be affected.

• kmeans_tidiers
• tidy.density
• tidy.spec

Unable to use

• tidy.ts: Doesn't preserve timestamps.
• Arima_tidiers: Not usable for autoplot.
• glm_tidiers: fortify is already defined in {ggplot2}. Considering to use autoplot internally for simplification.
• lm_tidiers: Same as glm.
• htest_tidiers: Plotting single test result is not very useful. Reconsider if there is good example.
• tidy.pairwise.htest: Doesn't contain confidence intervals. Considering to use {pairwiseCI}
• tidy.ftable: No need to use, simply call as.data.frame.

Maybe good to add a plotting parameter that allows you to turn off points to just show labels in factor plots if there isn't already a way to do this.

I have a survival grouping where the factor levels are different to alphabetical ordering.

levels(samplesGroups)
[1] "Low"  "High"

survfit keeps these in the correct order

> survivalFit
Call: survfit(formula = survivalData ~ samplesGroups)

	            n events median 0.95LCL 0.95UCL
samplesGroups=Low  17     11   3.63   1.838      NA
samplesGroups=High 17     10   4.55   0.992      NA

However, autoplot always places High first and Low second in the legend, because it ignores the ordering and just rearranges everything alphabetically. This should not happen.

Looks possible but can't work well with current ggmultiplot as legend must be considered as plot (grob). Needs separated class to dot this, or logic to merge the same legends when output.

p <- autoplot(lm(Petal.Width ~ Petal.Length, data = iris), colour = 'Species', data = iris)

grob <- ggplotGrob(p@plots[[1]])
leg <- grob$grobs[[which(grob$layout$name == "guide-box")]]

plots <- lapply(p@plots, function(x) x + theme(legend.position = 'none'))
plots[[5]] <- leg

layout_matrix = rbind(c(1, 2, 5), c(3, 4, 5))
do.call(grid.arrange, c(plots, list(ncol = 3, layout_matrix = layout_matrix))

Remove legacy code for ggplot2 v1.0 or prior once v1.1.0 has been released. Related to #47.

I firstly used the command

survivalPlot <- autoplot(survivalFit, conf.int = FALSE)

and got a sensible looking plot.

Then, I added a line colour and the placement of lines changed

autoplot(survivalFit, conf.int = FALSE, surv.colour = "red")

The legend also disappeared.

Hola,

Is it possible to use lda {MASS}? Can you please provide an example?

Thanks

Hi,
Love the package.

Maybe I'm just being stupid but can I use themes like I would in typical ggplot?

I'd hope
autoplot(lm(rnorm(10)~rnorm(10, 2))) +
theme_bw()

would theme the 4 facets with the black and white theme. This syntax works for autoplot in other cases (e.g. microbenchmark).

• Integrate ts model related classes as much
• Use geom_factory in label handling (#31)

In both install.packages and install_github.

Error in eval(expr, envir, enclos) : could not find function "eval"
Error : unable to load R code in package ‘ggfortify’
ERROR: lazy loading failed for package ‘ggfortify’

Maybe related to ggplto2 dev version:

• jrnold/ggthemes#36
• dkahle/ggmap#65

Unlike the base survival plot, it's not possible with ggfortify to set a different line colour for each patient group. Looking at the documentation of autoplot.survfit, it's only possible to choose a single colour for the parameter surv.colour.

plotdiag <-autoplot(result, which=c(1:3,5), mfrow=c(2,2))
plotdiag

***partdiage;

ggsave(plot=plotdiag, file='cuckoo-R-diag.png', h=4, w=6, units="in", dpi=300)

returns
Error in UseMethod("grid.draw") :
no applicable method for 'grid.draw' applied to an object of class "ggmultiplot"

I think Wickham modifies ggsave() to accept gtable objects. Do you create such an object? Perhaps add that as a class as well?

Using ggfortify 0.1.0 and ggplot 2.0.0

Many packages have a NEWS or README file with links on their CRAN webpage. ggfortify uses an unusual file format for these, so these files are not visible to the end user. Also, the NEWS file is missing information about what's new in the 0.3 version.

I'm trying to use autoplot on glm and lm objects. I am getting errors about dplyr exported objects.

> autoplot(lm(c(1:10)~ rnorm(10)))
Error: 'arrange_' is not an exported object from 'namespace:dplyr'

I tried to reinstall ggfortify from github and got
Error: processing vignette 'basics.Rmd' failed with diagnostics:
'select_' is not an exported object from 'namespace:dplyr'

Has something changed in dplyr that means these functions are broken? Thanks in advance for any help.

My sessionInfo is
R version 3.1.1 (2014-07-10)
Platform: x86_64-pc-linux-gnu (64-bit)

locale:
[1] LC_CTYPE=en_GB.UTF-8 LC_NUMERIC=C
[3] LC_TIME=en_GB.UTF-8 LC_COLLATE=en_GB.UTF-8
[5] LC_MONETARY=en_GB.UTF-8 LC_MESSAGES=en_GB.UTF-8
[7] LC_PAPER=en_GB.UTF-8 LC_NAME=C
[9] LC_ADDRESS=C LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_GB.UTF-8 LC_IDENTIFICATION=C

attached base packages:
[1] tcltk stats graphics grDevices utils datasets methods
[8] base

other attached packages:
[1] ggfortify_0.0.1 dplyr_0.2 sqldf_0.4-10 RSQLite_1.0.0
[5] DBI_0.3.1 gsubfn_0.6-6 proto_0.3-10 readxl_0.1.0
[9] ggplot2_1.0.0

#103

Decide how to handle ggrepel unique kwds.

• Prepare more vignette merging existing RPubs docs
• Release next version (Dec, 2015 or Jan, 2016 following CRAN policy?)
• Add link to vignette from existing RPubs.
• Rename VignetteIndexEntry

https://cran.r-project.org/web/checks/check_results_ggfortify.html

A while back, I wrote an autoplot method for spline basis objects: http://stats.stackexchange.com/a/179099/5736

Would you be interested in including this in ggfortify?

• Make post_fortify to always attach base_class
• Move ggplot instanciation to geom_factory, do not include duplicated data to geom.

By default, autoplot uses the names from survivalFit[["strata"]] as the legend levels. They have a variableName=value format. Programmer's variable names are not something that a figure to be published in a medical journal should have displayed. Perhaps autoplot should use code like gsub(".*=", '', strataNames) to place only the level name in the legend. autoplot also lacks any legend customisation options, such as the legend title.

Dear Sir,
I have written a function in which I used ggfortify to plot PCA the overall function is pretty simple, see below.
When I do the package (https://github.com/kendomaniac/docker4seq) check I get the following warning:
samplesPca: no visible global function definition for ‘autoplot’
Undefined global functions or variables:
autoplot

I import in my function ggfortify, but it seems that the autoplot in not imported from ggfortify
Could please suggest a way of fix this problem?
Cheers
Raf

######################################
#' @title Sample PCA on raw counts
#' @description This function runs PCA via ggfortify and ggplot2 on the output of samples2experiments function
#'
#' @param data, the counts dataframe generated by samples2experiemtn function
#' @param groups, a character string indicating association between samples and experimental groups
#' @param label, a boolean value, TRUE or FALSE to ad the names of the samples in the plot
#' @param label.size, an integer value indicating the size of label
#' @param output.file, a tab delimited file in which the first column are names from counts object from experiment.table.Rda file generated with samples2experiment function. In the pdf will be also rpesent the istigram fo the PCA variance dispersion.
#' @import ggfortify
#' @importFrom graphics plot
#' @importFrom grDevices pdf
#' @importFrom grDevices dev.off
#' @importFrom stats prcomp
#' @return a pdf file with the PCA plot
#' @examples
#'\dontrun{
#' load("test_experiment.tables.Rda")
#' samplesPca(data=counts, groups=c("d","eb","wb","ws"), label = TRUE, label.size = 3)
#' output.file="test.pdf", label = FALSE, label.size = 3)
#' }
#' @export
samplesPca <- function(data, groups=NULL, output.file=NULL, label, label.size){
if(!is.null(groups)){
data.df <- data.frame(names(data), as.factor(groups))
names(data.df) <- c("samples","groups")
if(is.null(output.file)){
autoplot.prcomp(prcomp(t(data)), data=data.df, colour='groups', label = label, label.size = label.size)
}else{
pdf(output.file)
plot(prcomp(t(data)), main="")
autoplot(prcomp(t(data)), data=data.df, colour='groups', label = label, label.size = label.size)
dev.off()
cat(paste("\nPCA is saved in ",output.file, sep=""))
}
}else{
if(is.null(output.file)){
autoplot(prcomp(t(data)))
}else{
pdf(output.file)
plot(prcomp(t(data)), main="")
autoplot(prcomp(t(data)))
dev.off()
cat(paste("\nPCA is saved in ",output.file, sep=""))
}
}
return(data.df)
}

If log is specified for fortify.survfit, the original values are returned unchanged.

fun <- switch(fun, log = function(x) x,

Is that a bug?

The legend is missing.

species <- as.numeric(factor(iris[, "Species"]))
autoplot(prcomp(iris[, 1:4]), colour = species, environment = environment())
Warning message:
In if (value %in% columns) { :
  the condition has length > 1 and only the first element will be used

Picking a column from the data frame works without a problem.

autoplot(prcomp(iris[, 1:4]), data = iris, colour = "Species") # Legend shown.

for example:

d <- rnorm(100)
i <- c(rep("A",50),rep("B",50))
fit <- kmeans(d,2,nstart=1000)
mds <- cmdscale(dist(d))
ggplot(as.data.frame(mds)) + geom_point(aes(mds[,1],mds[,2],shape=i),size=3)

So, I don't know how to use "i" by autoplot for different point...shape in autoplot is a logical value.

Hi,

I get similar error message as in #137 when doing any polynomial regression.
Please see reproducible example:

library(ggfortify)
model <- lm(mpg ~ poly(hp, 2, raw = TRUE),data = mtcars)
autoplot(model)
Error: Each variable must be a 1d atomic vector or list.
Problem variables: 'poly(hp, 2, raw = TRUE)'

par(mfrow = c(2,2))
plot(model)

binomial glm() models are often fit with a two column response variable. autoplot for these models is failing. Here is a reproducible example.

library(ggfortify)
data(menarche)
glm.out <- glm(cbind(Menarche, Total-Menarche) ~ Age, family=binomial(logit), data=menarche)
autoplot(glm.out) # fails 
#Error: Each variable must be a 1d atomic vector or list. 
#Problem variables: 'cbind(Menarche, Total - Menarche)'

# works fine
par(mfrow = c(2,2))
plot(glm.out)

Avoid to use rownames internally, follow dplyr 0.4- manner.

https://cran.r-project.org/web/packages/ggfortify/index.html shows:

Package ‘ggfortify’ was removed from the CRAN repository.

Formerly available versions can be obtained from the archive.

Archived on 2016-05-02 as check problems were not corrected despite reminders.

> install.packages fails as expected; however, the devtools install continues to work as expected after:

library(devtools)
install_github('sinhrks/ggfortify')

See #46.

Hello, thank you for the package. I find it really usefull. However, I would like the option to plot other axes instead of just the first two. Looking at the code I see that this is hardcoded. I could make a PR change that as an option if that's ok.

Should have the same interface with stats:::biplot.default.

Then refactor autoplot.pca_common.

Hello,

First and foremost, I would like to thank you and congratulate you on the effort you have put into this project. I was playing with the gglagplot function, when i noticed that it doesn't take a parameters to label the axis (y-axis - now get a "Data" label).

I was wondering if you can edit the list parameters, to accept ... (ellipsis), so that the parameters can be passed to ggplot function.

Thanks

Received kind email from CRAN:

Presumably from an update of one of

1 changepoint 2015-11-09
2 timeSeries 2015-11-09

we now see your package fail its tests with

checking tests ... ERROR
Running the tests in ‘tests/test-all.R’ failed.
Last 13 lines of output:
3: In princomp.default(df, center = TRUE, scale = FALSE) :
extra arguments 'center', 'scale' will be disregarded
4: In princomp.default(df, center = FALSE, scale = FALSE) :
extra arguments 'center', 'scale' will be disregarded
5: In princomp.default(USArrests, center = TRUE, scale = TRUE) :
extra arguments 'center', 'scale' will be disregarded
6: In princomp.default(USArrests, center = FALSE, scale = TRUE) :
extra arguments 'center', 'scale' will be disregarded
7: In princomp.default(USArrests, center = TRUE, scale = FALSE) :
extra arguments 'center', 'scale' will be disregarded
8: In princomp.default(USArrests, center = FALSE, scale = FALSE) :
extra arguments 'center', 'scale' will be disregarded
Execution halted