Simple program to make histograms like graphics.

[colt]

Hi, I need to do some histogram like graphics, but with different categories. For instance, in the "x" axis I want to put the name of the categories (like: car, airplane, boat, etc..) and in the y-axis the quantity (3,4,1, etc..). I would also like if each category could use different colors for each bar representing it. If these demands can be satisfied, the simpler the program is (easy to use) the better.

Any recommendations?

[george_toolan]

Try

yum install gnuplot

See http://www.google.de/search?q=gnuplot+histogram for details about gnuplot.

[Gareth Jones]

Spreadsheets like LibreOffice Calc, Gnumeric, and Calligra Sheets/KChart are probably the easiest GUI tools.

Gnuplot is great for more scientific/publication-quality graphs, and has the most flexible output options (various image formats, PostScript, PDF, SVG, EMF, various TeX/LaTeX/ConTeXt-formats). It does involve typing commands as text though, but this has the advantage of being easier to script if you want lots of similar plots.

There are dozens of others, some more specialized, some just front-ends to Gnuplot, but I'm not familiar with most of them: http://en.wikipedia.org/wiki/List_of_graphing_software.

[RupertPupkin]

I'd go with R, which is in the Fedora repos (yum install R). R is a command-line program but it's easy to use. After installing it just run 'R' without the quotes in a terminal window, which gives you a command prompt (>). From there you enter your commands, then enter quit("yes") (or just q("yes"), or the Ctrl-d key combo) when you want to exit. Here's a simple example of the commands to enter at the R command prompt to get a histogram with different colors per category:

categories <- c("car","airplane","boat","bicycle","skateboard")
quantities <- c(11,3,6,8,12)
barplot(quantities, names.arg=categories, col=rainbow(length(categories)),
        main="Transportation", xlab="Type", ylab="Count")

That barplot command should be all on one line (I just split it into two to make it easier to read here). By default the histogram will appear in a separate window. If you want to save it as an image file, say in a PNG file called myhistogram.png, then you'd use this command before running the barplot command:

png("myhistogram.png")

In this case the plot won't show up on the screen; you'll have to open the PNG file in an image viewer. Below is a picture of the histogram, along with a picture of the entire R session with the plot being saved as a PNG file in the current directory:

[colt]

Thanks all for the suggestions, was very useful!

[Adunaic]

I would agree with using R as well. It is brilliant for quick and clean plots. I would say the only real options are ROOT and R. ROOT is a bit of a pain so I would avoid it unless R cannot do what you want (i.e. the dataset gets to big for R to handle).

[j0534ng31]

RupertPupkin:

Wow! I was not aware that such tool (the R Project) existed.

Great info and a very clear/useful sample...

Thank you, sir!