Useful Thoughts and Actions Taken

Used it a lot at UFSC.  These are my notes on the usage of GnuPlot with examples.

Research Links

• GnuPlot 4.6 manual pdf
• You can do simple demo plots with Gnuplot to get accostomed.
• GnuPlot home page with links to SourceForge download
• Demo Scripts for GnuPlot
• Plot command options
• GnuPlot Examples
• Multi Axis Demo
• GnuPlot Tutorial
• ROW index in Excell – you are probably going to need this at some point  =ROW()
• GnuPlotting.org – many very good Example Scripts

 

You can download the file used in the example below and follow along: Processed_P1-Data.csv

Point Gnuplot to the correct directory to find the file full of data that you want to plot

The button with the folder icon and ChDir is what you want:

 

Set up GnuPlot to recognize comma as the field separator

set datafile separator "," 

This needs to match the field separator in your file.  My file has commas.  Watch out for excel files that export sometimes with only a CR and no LF character.  This will drive you crazy because it looks just fine when you look at the file with a text editor unless you are using something like NotePad++.  Then you will see it.  You need a right handed karate chopped n & r.  That would be \n\r.

 

Set the title of the plot

• set title "Title of the Plot"
• set xlabel "label"

 

Setting up X axis to be log scale

Use the following command to select a log scale on X:

set logscale x 10

When you need to switch back out of this mode:

unset logscale x

It appears they do not have an explicit command for "linear" mode.

 

Setting up the X and Y tics

Example: 

     set xtics 0,.5,10
     set xtics add ("Pi" 3.14159)

This will automatically generate tic marks every 0.5 along x, but will also add an explicit labeled tic mark at pi

Xtics

Setting up the X and Y minor tics

set mxtics 2   # this divides the interval between major tics by 2 and thus you get 1 tic in between major tics

 

Setting up axis formatting

Example:

           set format y "%4.0s"    – this gives 4 integer with nothing after the decimal point

format specifiers

 

Plotting with continuous lines

with line

Example:

plot "Current-Voltage-Power.csv" using 1:3 with line lt-1 lw 3,"Current-Voltage-Power.csv" using 1:7 with line lt-1 lw 3 axes x1y2,"Current-Voltage-Power.csv" using 1:6 with line lt-1 lw 3  

Gnuplot demo scripts

 

Setting the range to avoid taking log(0) and getting an error

When you try to plot a Log(X) axis you will see the following error

   x range must be greater than 0 for log scale

If you have not already done so set your X range so it does not include zero with:    set xrange[1e-10:1e-4] 

or whatever you like for the first value just make it >0.

 

Plotting Log(X) axis values correctly   Without this statement your X axis is labeled:  00000 00000 00000 .00001 .00010 etc and what you want is shown below.

If you use a log scale you will probably also want to format the log axis value labeling.

set logscale x 10
set format x '10^{%L}'  #<- enhanced text.

​

 

At this point the plot function should probably start working

plot "Processed_P1-Data.csv" using 4:8

 

If you want more than one plot on the same graph

In order to plot multiple lines in a single plot, simply put them in a single plot command like

plot 'AAA' u 1:2, 'BBB' u 1:2

 

If you want lines instead of only data points use:

plot "Processed_P1-Data.csv" using 2:3 with lines

Setting line color: You need this when GNUPLOT uses yellow as it is hard to see

set style line 6 linecolor rgb "blue"   

How to turn off the dataset legend that takes up so much room in the plot area

unset key

Set the plot size

set terminal wxt size 800,600 

Running a Script file

load "scriptname"

Plot Decimated data – 1 point for every X

• Large grouped data plotting
• Plotting data with GnuPlot – If you have a very, very large file with a lot of data points in it, it might be painfully slow to try to open the file in Excel in order to make a plot of a subset of the data. We have a 2.3GB file of XY points where we want to plot only a couple columns. In this case, gnuplot works really well.