[R] Best way to organize this data for plotting
Hello I have a number of data files which are outputs for each step of a simulation. The data is organized like this: dmindmax coef 5 6 0.490981 3 8 0.098056 5 6 0.425926 6 6 0.517860 2 4 0.527778 I would like to make a 3-d plot where one axis corresponds to (dmin, dmax) pair, one axis is the simulation step and the other is the mean of the coef column values that share the same (dmin, dmax). What's the easiest way to do this? I guess my question could be simplified to: what's the easiest way to get a list like (dmin1, dmax1) - mean(coef such that (dmin, dmax) = (dmin1, dmax1)) (dmin2, dmax2) - mean(coef such that (dmin, dmax) = (dmin2, dmax2)) ... (dminN, dmaxN) - mean(coef such that (dmin, dmax) = (dminN, dmaxN)) given a file like the one above? Thanks, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] contourplot help
Hello I'm using the following call to create a contourplot: library(lattice) m - as.matrix(read.table(data.txt)) contourplot(m[,3] ~ m[,2] * -m[,1], at = c(1e-6, 1e-5, 1e-4, 1e-3, 1e-2, 1e-1), scales = list(x = list(log = 10, labels = c(1, 10, 100), at = c(1, 10, 100)), y = list(labels = c(14, 12, 10, 8, 6, 4, 2)), at = c(-14, -12, -10, -8, -6, -4, -2)), labels = c(expression(10^-6), expression(10^-5), expression(10^-4), expression(10^-3), expression(10^-2), expression(10^-1), expression(10^0)), xlim = c(0.75, 10^2), xlab = Out-degree, ylab = In-degree) Which gives the the output in the file below http://ompldr.org/vNm4xag/contour.eps As it can be seen, the level labels are not displayed nicely because there's not enough room for them. Also, the 10^-1 label is not displayed. Is there a way for me to hardcode the position of each label? I tried setting labels = F and then calling text() for each one, but that doesn't work. If that's not possible, one option would be to color each level line differently and then add a legend. Is it possible to do that? Finally, how can I remove the tick marks from the top and right axes? Thanks, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Working with tables with missing levels
Hello I'm trying to write a function to calculate the relative entropy between two distributions. The data I have is in table format, for example: t1 - prop.table(table(c(0,0,2,4,4))) t2 - prop.table(table(c(0,2,2,2,3))) t1 0 2 4 0.4 0.2 0.4 t2 0 2 3 0.2 0.6 0.2 The relative entropy is given by H[P||Q] = sum(p * log2(p/q)) with the conventions that 0*log2(0/q) = 0 and p*log2(p/0) = Inf. I'm not sure about what is the best way to achieve that. Is there a way to test if a table has a value for a given level, so that I can detect that, for example, t1 is missing levels 1 and 3 and t2 is missing levels 1 and 4 (is level the correct terminology here?)? Simply trying to access t1[[1]], for example, gives a subscript out of bounds error. Another option would be to expand the tables, so that, for example, t1 becomes 0 1 2 3 4 0.4 0.0 0.2 0.0 0.4 Is there a way to do that? Thanks, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Working with tables with missing levels
On Mon, 2009-07-27 at 16:34 -0300, Henrique Dallazuanna wrote:
Try this:
t1 - prop.table(table(factor(c(0,0,2,4,4), levels = 0:4)))
t2 - prop.table(table(factor(c(0,2,2,2,3), levels = 0:4)))
Is there a way to do this given an already existing table? The problem
is that I actually build the distributions as I read data from files,
something like
distr - NULL
for (file in files) {
x - as.matrix(read.table(file))
t - c(distr, table(x))
distr - tapply(t, names(t), sum)
}
distr - prop.table(distr)
So I only know the maximum level after the distributions are created.
Thanks,
Andre
__
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
[R] Postcript font size
Hello I'm doing a number of plots and in all of them I'm specifying the same font size. However, comparing one plot to the other, they have fonts of different sizes, so it appears some scaling is being done. I tried using both postcript(pointsize = ...) and par(ps = ...), but the results were the same. Is there a way to specify a font size that is consistent among calls to plot()? Thanks, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Postcript font size
On Thu, 2009-06-11 at 19:09 -0300, Andre Nathan wrote: Is there a way to specify a font size that is consistent among calls to plot()? To put it correctly: is there a way to specify a font size that is consistent among calls to postcript()? All the fonts in the same file have the same size, but comparing files, even when par(ps = ...) or postcript(pointsize = ...) are specified the same way, the fonts have different sizes. Best regards, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Postcript font size
On Thu, 2009-06-11 at 23:48 +0100, ted.hard...@manchester.ac.uk wrote: How are you comparing? A: Looking into the PostScript code in the files, and identifying the font-sizes where they are set in the code? This. To get the size I wanted in both files, I had to use par(ps = 15) in one of them, and par(ps = 18) in the other. Strangely, the generated files had the following PS code: /ps 12 def /Font1 findfont 12 s I don't know PostScript, but I assume this means pointsize = 12. Best, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Axis label spanning multiple plots
Hello On Wed, 2009-05-27 at 13:38 -0600, Greg Snow wrote: Create an outer margin (see ?par), then use mtext to put the title in the outer margin. Sorry for taking that long to reply... I created an outer margin with par(oma = c (0, 2, 0, 0)) and then did par(mfrow = c(2, 1)) plot(...) plot(...) mtext(title 1, side = 2, outer = T) With that, the text title 1 appears to the left of the second plot, and not to the left of both plots, centered between them, which is what I was trying to do. Is there a way to accomplish that (without manual text positioning, that is)? Thanks, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Axis label spanning multiple plots
Hello I need to plot 3 graphs in a single column; the top two plots have the same title, and I would like it to be written only once, centered horizontally and spanning the two plots. Something like t ++ || i || || t ++ l ++ || e || || 1 ++ t ++ i || t || l || e || 2 ++ x-title Is there a parameter which allows me to do that automatically, or should I use text() and position the title manually? Thanks in advance, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Multiple plot margins
On Wed, 2009-05-13 at 11:22 +0200, Uwe Ligges wrote: If not, example: par(mfrow = c(2,3), mar = c(0,0,0,0), oma = c(5,5,0,0), xpd=NA) plot(1, xaxt=n, xlab=, ylab=A) plot(1, xaxt=n, yaxt=n, xlab=, ylab=) plot(1, xaxt=n, yaxt=n, xlab=, ylab=) plot(1, xlab=I, ylab=B) plot(1, xlab=II, ylab=, yaxt=n) plot(1, xlab=III, ylab=, yaxt=n) Thank you. I don't know what I did wrong, but that worked. Best regards, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Multiple plot margins
Hello I'm plotting 6 graphs using mfrow = c(2, 3). In these plots, only graphs in the first column have titles for the y axis, and only the ones in the last row have titles for the x axis. I'd like all plots to be of the same size, and I'm trying to keep them as near each other as possible, but I'm having the following problem. If I make a single call to par(mar = ...), to leave room on the left and bottom for the axes titles, a lot of space will be wasted because not all graphs need titles; however, if I make one call of par(mar = ...) per plot, to have finer control of the margins, the first column and last row plots will be smaller than the rest, because the titles use up some of their space. I thought that setting large enough values for oma would do what I want, but it doesn't appear to work if mar is too small. To illustrate better what I'm trying to do: l +-+ +-+ +-+ a | | | | | | b | | | | | | e | | | | | | l +-+ +-+ +-+ l +-+ +-+ +-+ a | | | | | | b | | | | | | e | | | | | | l +-+ +-+ +-+ label label label where the margins between each plot should be narrow. Should I just plot the graphs without axis titles and then use text() to manually position them? Thanks in advance, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Incrementally building histograms
Hello
I need to build a histogram from data (numbers in the [0,1] interval)
stored in a number of different files. The total amount of data is very
large, so I can't load everything to memory and then simply call hist().
Since what I actually need are the histogram counts, I'm currently doing
it like this:
breaks - seq(0, 1, by = 0.01)
files - list.files(pattern = some pattern)
counts - 0
for (file in files) {
data - scan(file, quiet = T)
h - hist(data, plot = F, breaks = breaks)
counts - counts + h$counts
}
# and then work with `counts' here
Is there a more efficient and/or idiomatic way to do this?
Thanks,
Andre
__
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Re: [R] Combining tables
Thanks Jim, it worked great! On Sun, 2008-09-14 at 21:27 -0400, jim holtman wrote: try this: c(tx,ty) 1 2 3 3 4 3 2 1 4 1 z - c(tx,ty) tapply(z, names(z), sum) 1 2 3 4 3 2 5 1 __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Combining tables
Hello Say I have the following data, and it's distribution given by table(): x - c(1, 1, 1, 2, 2, 3) tx - table(x) tx x 1 2 3 3 2 1 Now say I have new data, y - c(3, 3, 3, 3, 4) ty - table(y) ty y 3 4 4 1 Is there a way to combine tx and ty in such a way to give me the distribution below? 1 2 3 4 3 2 5 1 Essentially what I'm looking for is something equivalent to table(c(x,y)), but x and y are too large and I'd like to avoid the concatenation. Thanks in advance, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] R rpm for Mandriva 2008.1 ?
Hi Joanne There is a package in Mandriva's repository, called R-base. You can isntall it with urpmi R-base. HTH, Andre On Sat, 2008-09-13 at 16:48 +0100, Joanne Demmler wrote: Are there any R rpm's for Mandriva 2008.1? I found a couple of dodgy ones that wouldn't install so far (rpmfind.net). (Sorry that seems so far the only Linux distribution that recognizes my hardware on my new laptop correctly. I'm more familiar to Fedora and Suse so perhaps I'm just doing something wrong?) Joanne __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Histograms without bars
Hello I'd like to plot a histogram of some data composed of real numbers. The bin width I'm using is ~ 0.01, which results in high values in the y axis, so that the area under each bar corresponds to the probability of the data in that range. Is is possible to plot points whose y coordinate correspond to that probability, instead of plotting the histogram bars? In other words, instead of having a bar of width 0.01 and height, say, 80, I'd like to have a single point at y = 0.8. I know I could use the values returned from hist() and calculate that manually, but maybe there's a better way to do that. I've been using prop.table() to plot these histograms for integer numbers, but now I need binning and I'm not sure what is the best way to proceed. Thanks in advance, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Multiple plots question
Hello (Sorry if this appears twice, had some mail problems...) I have a number of different data sets, each loaded as a matrix. I'd like to plot them in a way that the data in the first column of each matrix is plotted on the same pair of axes. What I'm doing now is to call plot() for the data on the first matrix, then call points() for the other ones. However, the axes are set by R according to the data passed to plot(), and sometimes the data passed to points() has larger values on the x axis, and the plot ends up being cut (the y axis is not a problem since they're all probabilities). Is there a way to dynamically adapt the axes so that all data can be seen? I know I could build a new matrix with the columns I want to plot but each matrix has 1 million rows, so I figure this would be inefficient. Do I have to check beforehand which column has the largest value and call plot() on it, and then points() on the others, or is there an automatic way to do this? Thanks in advance, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Multiple plots question
Hello I have a number of different data sets, each loaded as a matrix. I'd like to plot them in a way that the data in the first column of each matrix is plotted on the same pair of axes. What I'm doing now is to call plot() for the data on the first matrix, then call points() for the other ones. However, the axes are set by R according to the data passed to plot(), and sometimes the data passed to points() has larger values on the x axis, and the plot ends up being cut (the y axis is not a problem since they're all probabilities). Is there a way to dynamically adapt the axes so that all data can be seen? I know I could build a new matrix with the columns I want to plot but each matrix has 1 million rows, so I figure this would be inefficient. Do I have to check beforehand which column has the largest value and call plot() on it, and then points() on the others, or is there an automatic way to do this? Thanks in advance, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Raw histogram plots
On Wed, 2008-02-27 at 14:15 +1300, Peter Alspach wrote: If I understand you correctly, you could try a barplot() on the result of table(). Hmm, table() does the counting exactly the way I want, i.e., just counting individual values. Is there a way to extract the counts vs. the values from a table, so that I can pass them as the x and y arguments to plot()? Thanks, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Raw histogram plots
On Wed, 2008-02-27 at 08:48 -0500, Charilaos Skiadas wrote: x - table(rbinom(20,2,0.5)) plot(names(x),x) should do it. You can also try just plot(x). Use prop.table on table if you want the relative frequencies instead. Yes, names is what I needed :) Thanks for the prop.table hint. I looked everywhere but none of my searches hinted at table/table.prop. You guys' help has been invaluable for me. Thanks again, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Raw histogram plots
Hello I need to plot a histogram, but insted of using bars, I'd like to plot the data points. I've been doing it like this so far: h - hist(x, plot = F) plot(y = x$counts / sum(x$counts), x = x$breaks[2:length(x$breaks)], type = p, log = xy) Sometimes I want to have a look at the raw data (avoiding any kind of binning). When x only contains integers, it's easy to just use bins of size 1 when generating h with breaks = seq(0, max(x)). Is there any way to do something similar when x consists of fractional data? What I'm doing is setting a small bin length (for example, breaks = seq(0, 1, by = 1e-6), but there's still a chance that points will be grouped in a single bin. Is there a better way to do this kind of raw histogram plotting? Thanks, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Raw histogram plots
I know about stem, but the data set has 1 million points, so it's not very useful here. I want to avoid binning just to have an idea about the shape of the distribution, before deciding how I'll bin it. Andre On Tue, 2008-02-26 at 16:20 -0600, roger koenker wrote: take a look at ?stem There is still a place for handtools in the age of integrated circuits. Of course, avoiding binning isn't really desirable. url:www.econ.uiuc.edu/~rogerRoger Koenker email[EMAIL PROTECTED]Department of Economics vox: 217-333-4558University of Illinois fax: 217-244-6678Champaign, IL 61820 On Feb 26, 2008, at 4:10 PM, Andre Nathan wrote: Hello I need to plot a histogram, but insted of using bars, I'd like to plot the data points. I've been doing it like this so far: h - hist(x, plot = F) plot(y = x$counts / sum(x$counts), x = x$breaks[2:length(x$breaks)], type = p, log = xy) Sometimes I want to have a look at the raw data (avoiding any kind of binning). When x only contains integers, it's easy to just use bins of size 1 when generating h with breaks = seq(0, max(x)). Is there any way to do something similar when x consists of fractional data? What I'm doing is setting a small bin length (for example, breaks = seq(0, 1, by = 1e-6), but there's still a chance that points will be grouped in a single bin. Is there a better way to do this kind of raw histogram plotting? Thanks, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] A more idiomatic way to write this
Hello,
I have a vector of 1,000,000 numbers and another vector of 1,000
divisors. What I'd like to do is to divide the first 1,000 numbers of
the first vector by the first divisor, then the next 1,000 by the second
divisor and so on. I came up with this, but I was wondering if there is
a more idiomatic, R-like way to write it:
x - ...
divs - ...
for (i in seq(from = 1, to = 100, by = 1000)) {
x[i:(i - 1 + 1000)] - x[i:(i - 1 + 1000)] / divs[i %/% 1000 + 1]
}
Any suggestions are welcome.
Thanks in advance,
Andre
__
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Re: [R] A more idiomatic way to write this
On Mon, 2008-02-25 at 11:52 +1000, [EMAIL PROTECTED] wrote: x - x/rep(divs, each = 1000) Fantastic :) Thanks Bill, Andrew and Rolf Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Geometric progression
On Sun, 2008-02-24 at 23:26 -0300, Andre Nathan wrote: gp(1, 2, 10) [1]1248 16 32 64 128 256 512 1024 Actually, [1]1248 16 32 64 128 256 512 Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Geometric progression
On Sun, 2008-02-24 at 21:39 -0500, Charilaos Skiadas wrote: 2^(0:9) I guess it's so simple that I'd never think of that... Thanks! Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Geometric progression
Hi
I'm pretty sure there's some built-in function to do the equivalent of
this, but I couldn't find one anywhere:
gp - function(init, mult, n)
{
if (n == 1)
init
else
pg(c(init, init[length(init)] * mult), mult, n-1)
}
gp(1, 2, 10)
[1]1248 16 32 64 128 256 512 1024
Any pointers?
Thanks,
Andre
__
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Re: [R] How to join path with arguments
Hyunchul Kim wrote: How to format and join strings ? For example, like following short python examples. * name1 = 'sample-plot' filename = '%s.png' % name1 Two options are sprintf() and paste(): name1 = sample-plot sprintf(%s.png, name1) [1] sample-plot.png paste(name1, png, sep = .) [1] sample-plot.png HTH, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Questions about histograms
Hello I'm doing some experiments with the various histogram functions and I have a two questions about the prob option and binning. First, here's a simple plot of my data using the default hist() function: hist(data[,1], prob = TRUE, xlim = c(0, 35)) http://go.sneakymustard.com/tmp/hist.jpg My first question is regarding the resulting plot from hist.scott() and hist.FD(), from the MASS package. I'm setting prob to TRUE in these functions, but as it can be seen in the images below, the value for the first bar of the histogram is well above 1.0. Shouldn't the total area be 1.0 in the case of prob = TRUE? hist.scott(data[,1], prob = TRUE, xlim=c(0, 35)) http://go.sneakymustard.com/tmp/scott.jpg hist.FD(data[,1], prob = TRUE, xlim=c(0, 35)) http://go.sneakymustard.com/tmp/FD.jpg Is there anything I can do to fix these plots? My second question is related to binning. Is there a function or package that allows one to use logarithmic binning in R, that is, create bins such that the length of a bin is a multiple of the length of the one before it? Pointers to the appropriate docs are welcome, I've been searching for this and couldn't find any info. Best regards, Andre __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
