gnuplot-info

[Gnuplot-info] automatic binning for histogram

[Robbie M. <ro...@ac...>]

Hello all

Say I have a single column of sorted numbers (given
here horizontally):

  3 3 4 4 4 5 6 6 8

and want to plot a histogram, something like:

    |
    |
    |               =
    |           =   =       =
    |           =   =   =   =       =
    +---+---+---+---+---+---+---+---+---
    0   1   2   3   4   5   6   7   8

In order to get the data into a suitable state for
plotting with 'histogram', do I have to fill the bins
myself, or can 'gnuplot' do that for me?

If this is not supported, would it be a useful
addition?  The case above is integer-valued and the bin
width is intuitively one.  More complicated
integer-valued data and continuous data would need the
used to define the bin width, I suppose.

I couldn't see anything in the PDF documentation or in
Janert (2010).

any suggestions gratefully received, Robbie
---
Robbie Morrison
PhD student -- policy-oriented energy system simulation
Technical University of Berlin (TU-Berlin), Germany
University email (redirected) : mor...@ie...
Webmail (preferred)           : ro...@ac...
[from Webmail client]

Re: [Gnuplot-info] automatic binning for histogram

[MadMarvin <ja...@th...>]

Hi Robbie,
not, this is not possible, at least not without making your life very
difficult. Gnuplot is not and never was meant to be a calculation tool,
that's what OOCalc and Excel are for.

Besides, you will always have to adjust your data carefully, especially in
scientific applications, i. e. when trying to fit a distribution function to
your data or choosing the binning size. Even Excel only uses the sqrt(n)
rule although more sophisticade approaches are available:
http://en.wikipedia.org/wiki/Histogram#Number_of_bins_and_width

I suggest using the COUNTIF functions of the above mentioned spreadsheet
tools instead.
-- 
View this message in context: http://old.nabble.com/automatic-binning-for-histogram-tp32038612p32052193.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] automatic binning for histogram

[Gareth W. <g.w...@sh...>]

On 11 Jul 2011, at 16:42, Robbie Morrison wrote:

> Say I have a single column of sorted numbers (given
> here horizontally):
> 
>  3 3 4 4 4 5 6 6 8
> 
> and want to plot a histogram, something like:
> 
>    |
>    |
>    |               =
>    |           =   =       =
>    |           =   =   =   =       =
>    +---+---+---+---+---+---+---+---+---
>    0   1   2   3   4   5   6   7   8
> 
> In order to get the data into a suitable state for
> plotting with 'histogram', do I have to fill the bins
> myself, or can 'gnuplot' do that for me?

You could try this to start with.  It assumes you have your raw data in a single-column file called foo.dat.

     set boxwidth 1
     bw = 1
     bin(x,width)=width*floor(x/width) + bw/2.0
     set xrange [0:*]
     set yrange [0:*]
     plot 'foo.dat' using (bin($1,bw)):(1.0) smooth freq with boxes

Best wishes,

Gareth