I was reading through this blog post on R-bloggers and I'm confused by the last section of the code and can't figure it out.
http://www.r-bloggers.com/self-organising-maps-for-customer-segmentation-using-r/
I've attempted to recreate this with my own data. I have 5 variables that follow an exponential distribution with 2755 points.
I am fine with and can plot the map that it generates:
plot(som_model, type="codes")
The section of the code I don't understand is the:
var <- 1
var_unscaled <- aggregate(as.numeric(training[,var]),by=list(som_model$unit.classif),FUN = mean, simplify=TRUE)[,2]
plot(som_model, type = "property", property=var_unscaled, main = names(training)[var], palette.name=coolBlueHotRed)
As I understand it, this section of the code is suppose to be plotting one of the variables over the map to see what it looks like but this is where I run into problems. When I run this section of the code I get the warning:
Warning message:
In bgcolors[!is.na(showcolors)] <- bgcol[showcolors[!is.na(showcolors)]] :
number of items to replace is not a multiple of replacement length
and it produces the plot:
Which just some how doesn't look right...
Now what I think it has come down to is the way the aggregate function has re-ordered the data. The length of var_unscaled is 789 and the length of som_model$data, training[,var] and unit.classif are all of length 2755. I tried plotting the aggregated data, the result was no warning but an unintelligible graph (as expected).
Now I think it has done this because unit.classif has a lot of repeated numbers inside it and that's why it has reduced in size.
The question is, do I worry about the warning? Is it producing an accurate graph? What exactly is the "Property"'s section looking for in the plot command? Is there a different way I could "Aggregate" the data?
I think that you have to create the palette color. If you put the argument
coolBlueHotRed <- function(n, alpha = 1) {rainbow(n, end=4/6, alpha=alpha)[n:1]}
and then try to get a plot, for example
plot(som_model, type = "count", palette.name = coolBlueHotRed)
the end is succesful.
This link can help you: http://rgm3.lab.nig.ac.jp/RGM/R_rdfile?f=kohonen/man/plot.kohonen.Rd&d=R_CC
I think that not all of the cells on your map have points inside.
You have 30 by 30 map and about 2700 points. In average it's about 3 points per cell. With high probability some cells have more than 3 points and some cells are empty.
The code in the post on R-bloggers works well when all of the cells have points inside.
To make it work on your data try change this part:
var <- 1
var_unscaled <- aggregate(as.numeric(training[, var]), by = list(som_model$unit.classif), FUN = mean, simplify = TRUE)[, 2]
plot(som_model, type = "property", property = var_unscaled, main = names(training)[var], palette.name = coolBlueHotRed)
with this one:
var <- 1
var_unscaled <- aggregate(as.numeric(data.temp[, data.classes][, var]),
by = list(som_model$unit.classif),
FUN = mean,
simplify = T)
v_u <- rep(0, max(var_unscaled$Group.1))
v_u[var_unscaled$Group.1] <- var_unscaled$x
plot(som_model,
type = "property",
property = v_u,
main = colnames(data.temp[, data.classes])[var],
palette.name = coolBlueHotRed)
Hope it helps.
Just add these functions to your script:
coolBlueHotRed <- function(n, alpha = 1) {rainbow(n, end=4/6, alpha=alpha)[n:1]}
pretty_palette <- c("#1f77b4","#ff7f0e","#2ca02c", "#d62728","#9467bd","#8c564b","#e377c2")
Related
I am trying to turn off the display of plot in R.
I read Disable GUI, graphics devices in R but the only solution given is to write the plot to a file.
What if I don't want to pollute the workspace and what if I don't have write permission ?
I tried options(device=NULL) but it didn't work.
The context is the package NbClust : I want what NbClust() returns but I do not want to display the plot it does.
Thanks in advance !
edit : Here is a reproducible example using data from the rattle package :)
data(wine, package="rattle")
df <- scale (wine[-1])
library(NbClust)
# This produces a graph output which I don't want
nc <- NbClust(df, min.nc=2, max.nc=15, method="kmeans")
# This is the plot I want ;)
barplot(table(nc$Best.n[1,]),
xlab="Numer of Clusters", ylab="Number of Criteria",
main="Number of Clusters Chosen by 26 Criteria")
You can wrap the call in
pdf(file = NULL)
and
dev.off()
This sends all the output to a null file which effectively hides it.
Luckily it seems that NbClust is one giant messy function with some other functions in it and lots of icky looking code. The plotting is done in one of two places.
Create a copy of NbClust:
> MyNbClust = NbClust
and then edit this function. Change the header to:
MyNbClust <-
function (data, diss = "NULL", distance = "euclidean", min.nc = 2,
max.nc = 15, method = "ward", index = "all", alphaBeale = 0.1, plotetc=FALSE)
{
and then wrap the plotting code in if blocks. Around line 1588:
if(plotetc){
par(mfrow = c(1, 2))
[etc]
cat(paste(...
}
and similarly around line 1610. Save. Now use:
nc = MyNbClust(...etc....)
and you see no plots unless you add plotetc=TRUE.
Then ask the devs to include your patch.
I am trying to plot FEVD (forecast error variance decomposition) for my VAR analysis. As you can see on the image, the legend screws up the graph and information. as this is an automatically created legend, I don’t know how to reposition it. I do not know much yet about plotting in R.
The only code i use to get this is :
library(vars)
var <- VAR(varTable2 , p=4 , type = "both")
plot(fevd(var, n.ahead = 10 ))
Thanks in advance
Legends do not resize well in R. You have to set your plotting window first and then chart your data.
Here's how to do it in Windows. win.graph opens a blank plotting window of the specified width. In Unix/Linux, you should look at X11() and in Mac, at quartz(). You might also consider shorter variable names.
library(vars)
data(Canada)
colnames(Canada) <-c("Long column name1","Long column name2","Long column name3","Long column name4")
var <- VAR(Canada , p=4 , type = "both")
win.graph(width=13,height=8)
plot(fevd(var, n.ahead = 10 ))
I'm visualizing a data set with the heatmap.2 function from the gplots package in R. Basically I'm performing a hierarchical clustering analysis on the original data, while forcing the heatmap to display a limited version of the data (between -3 and +3) to limit the effect of outliers on the appearance of the heatmap, while still retaining the original clustering. When I use the full data set (fullmousedatamat), it works just fine. However, when I use a partial data set (partialmousedatamat), and want to plot it using the same key as the full data set, a couple colors are dropped out of the key and I can't figure out why.
Here is a gist containing the relevant data sets and analyses:
https://gist.github.com/jeffbruce/7412f567ac57fe1721a3
Notice how the 4th color on either side of the centre white color are dropped out. This feels like a bug to me maybe. I get the following warning message which I'm not sure how to interpret:
Warning message:
In image.default(z = matrix(z, ncol = 1), col = col, breaks = tmpbreaks, :
unsorted 'breaks' will be sorted before use
Thanks for your help!
I came across the same issue and I had to go through the code for heatmap.2 to figure it out.
It turns out that symkey=T, which is the default, adds the extreme values of the data at both ends of breaks, rendering it un-sorted:
tmpbreaks <- breaks
if (symkey) {
max.raw <- max(abs(c(x, breaks)), na.rm = TRUE)
min.raw <- -max.raw
tmpbreaks[1] <- -max(abs(x), na.rm = TRUE)
tmpbreaks[length(tmpbreaks)] <- max(abs(x), na.rm = TRUE)
}
Therefore, the simple way to solve this is adding symkey=F if you are providing your own breaks.
I have created a mosaic plot using mosaic function in the vcd package. Now I wish to add some annotations using labeling_cells. Unfortunately, I get an error. The problem might be that it is not the standard Titanic example...
library("grid"); library("vcd")
dataset <- read.table("http://bit.ly/1aJTI1C")
# prepare data for plot as a "Structured Contingency Table"
data1 <- structable(dauer ~ groesse + ort, dataset)
# basic plot
mosaic(data1,
# separate the two elements of the plot
split_vertical = c(T, T, F),
# put the names in the right places and adds boxes
labeling_args = list(tl_labels = TRUE,
tl_varnames = FALSE,
boxes = TRUE),
# grip remains open
pop=FALSE
)
# structure that matches plot, but it does not help
#match<-t(data1)
# try to add labels
labeling_cells(text = data1, clip = FALSE)(data1)
This results in:
# Error in ifelse(abbreviate_varnames, sapply(seq_along(dn), function(i) abbreviate(dn[i], :
# replacement has length zero
# In addition: Warning message:
# In rep(no, length.out = length(ans)) :
# 'x' is NULL so the result will be NULL
Another problem I have is that the boxes do not fit the labels. If you have a hint for that just let me know as well!
It's my first question here, so please excuse potential errors!
Thanks a lot!
Fixed upstream in vcd 1.4-4, but note that you can simply use
mosaic(data1, labeling = labeling_values)
Yes, this is quite confusing and ought to be fixed in labeling_cells(). For some reason the data in the labeling should be a regular table, not a structable. I'll raise this with David, the principal author of mosaic() and package maintainer.
If you know it it's easy to work around it, though:
labeling_cells(text = as.table(data1), clip = FALSE)(as.table(data1))
I need to draw lines from the data stored in a text file.
So far I am able only to draw points on a graph and i would like to have them as lines (line graph).
Here's the code:
pupil_data <- read.table("C:/a1t_left_test.dat", header=T, sep="\t")
max_y <- max(pupil_data$PupilLeft)
plot(NA,NA,xlim=c(0,length(pupil_data$PupilLeft)), ylim=c(2,max_y));
for (i in 1:(length(pupil_data$PupilLeft) - 1))
{
points(i, y = pupil_data$PupilLeft[i], type = "o", col = "red", cex = 0.5, lwd = 2.0)
}
Please help me change this line of code:
points(i, y = pupil_data$PupilLeft[i], type = "o", col = "red")
to draw lines from the data.
Here is the data in the file:
PupilLeft
3.553479
3.539469
3.527239
3.613131
3.649437
3.632779
3.614373
3.605981
3.595985
3.630766
3.590724
3.626535
3.62386
3.619688
3.595711
3.627841
3.623596
3.650569
3.64876
By default, R will plot a single vector as the y coordinates, and use a sequence for the x coordinates. So to make the plot you are after, all you need is:
plot(pupil_data$PupilLeft, type = "o")
You haven't provided any example data, but you can see this with the built-in iris data set:
plot(iris[,1], type = "o")
This does in fact plot the points as lines. If you are actually getting points without lines, you'll need to provide a working example with your data to figure out why.
EDIT:
Your original code doesn't work because of the loop. You are in effect asking R to plot a line connecting a single point to itself each time through the loop. The next time through the loop R doesn't know that there are other points that you want connected; if it did, this would break the intended use of points, which is to add points/lines to an existing plot.
Of course, the line connecting a point to itself doesn't really make sense, and so it isn't plotted (or is plotted too small to see, same result).
Your example is most easily done without a loop:
PupilLeft <- c(3.553479 ,3.539469 ,3.527239 ,3.613131 ,3.649437 ,3.632779 ,3.614373
,3.605981 ,3.595985 ,3.630766 ,3.590724 ,3.626535 ,3.62386 ,3.619688
,3.595711 ,3.627841 ,3.623596 ,3.650569 ,3.64876)
plot(PupilLeft, type = 'o')
If you really do need to use a loop, then the coding becomes more involved. One approach would be to use a closure:
makeaddpoint <- function(firstpoint){
## firstpoint is the y value of the first point in the series
lastpt <- firstpoint
lastptind <- 1
addpoint <- function(nextpt, ...){
pts <- rbind(c(lastptind, lastpt), c(lastptind + 1, nextpt))
points(pts, ... )
lastpt <<- nextpt
lastptind <<- lastptind + 1
}
return(addpoint)
}
myaddpoint <- makeaddpoint(PupilLeft[1])
plot(NA,NA,xlim=c(0,length(PupilLeft)), ylim=c(2,max(PupilLeft)))
for (i in 2:(length(PupilLeft)))
{
myaddpoint(PupilLeft[i], type = "o")
}
You can then wrap the myaddpoint call in the for loop with whatever testing you need to decide whether or not you will actually plot that point. The function returned by makeaddpoint will keep track of the plot indexing for you.
This is normal programming for Lisp-like languages. If you find it confusing you can do this without a closure, but you'll need to handle incrementing the index and storing the previous point value 'manually' in your loop.
There is a strong aversion among experienced R coders to using for-loops when not really needed. This is an example of a loop-less use of a vectorized function named segments that takes 4 vectors as arguments: x0,y0, x1,y1
npups <-length(pupil_data$PupilLeft)
segments(1:(npups-1), pupil_data$PupilLeft[-npups], # the starting points
2:npups, pupil_data$PupilLeft[-1] ) # the ending points