pareto function

[David Bateman]

Here is a patch that adds the pareto function. At the same time it fixes
two bugs in plotyy and __go_draw_axes__ that prevented pareto from
working correctly.

D.

Index: scripts/plot/__go_draw_axes__.m
===================================================================
RCS file: /usr/local/cvsroot/octave/scripts/plot/__go_draw_axes__.m,v
retrieving revision 1.77
diff -c -r1.77 __go_draw_axes__.m
*** scripts/plot/__go_draw_axes__.m     29 Nov 2007 01:57:11 -0000      1.77
--- scripts/plot/__go_draw_axes__.m     29 Nov 2007 22:55:04 -0000
***************
*** 1590,1601 ****
        nlabels = numel (labels);
        fprintf (plot_stream, "set format %s \"%%s\";\n", ax);
        if (mirror)
!         fprintf (plot_stream, "set %stics %s (", ax, colorspec);
        else
!         fprintf (plot_stream, "set %stics nomirror %s (", ax, colorspec);
        endif
        for i = 1:ntics
!         fprintf (plot_stream, " \"%s\" %g", labels(k++), tics(i))
          if (i < ntics)
            fputs (plot_stream, ", ");
          endif
--- 1590,1602 ----
        nlabels = numel (labels);
        fprintf (plot_stream, "set format %s \"%%s\";\n", ax);
        if (mirror)
!         fprintf (plot_stream, "set %stics (", ax);
        else
!         fprintf (plot_stream, "set %stics nomirror (", ax);
        endif
        for i = 1:ntics
!         fprintf (plot_stream, " \"%s\" %g", 
!                  regexprep (labels(k++), "%", "%%"), tics(i))
          if (i < ntics)
            fputs (plot_stream, ", ");
          endif
***************
*** 1603,1609 ****
            k = 1;
          endif
        endfor
!       fputs (plot_stream, ");\n");
        else
        error ("unsupported type of ticklabel");
        endif
--- 1604,1610 ----
            k = 1;
          endif
        endfor
!       fprintf (plot_stream, ") %s;\n", colorspec);
        else
        error ("unsupported type of ticklabel");
        endif
Index: scripts/plot/plotyy.m
===================================================================
RCS file: /usr/local/cvsroot/octave/scripts/plot/plotyy.m,v
retrieving revision 1.6
diff -c -r1.6 plotyy.m
*** scripts/plot/plotyy.m       29 Nov 2007 19:07:29 -0000      1.6
--- scripts/plot/plotyy.m       29 Nov 2007 22:55:04 -0000
***************
*** 93,99 ****
    xlim = [min([x1(:); x2(:)]), max([x1(:); x2(:)])];
  
    h1 = feval (fun1, x1, y1);
!   set (ax(1), "ycolor", get (h1(1), "color"));
    set (ax(1), "position", get (ax(1), "outerposition"));
    set (ax(1), "xlim", xlim);
  
--- 93,100 ----
    xlim = [min([x1(:); x2(:)]), max([x1(:); x2(:)])];
  
    h1 = feval (fun1, x1, y1);
! 
!   set (ax(1), "ycolor", getcolor (h1(1)));
    set (ax(1), "position", get (ax(1), "outerposition"));
    set (ax(1), "xlim", xlim);
  
***************
*** 105,111 ****
  
    h2 = feval (fun2, x2, y2);
    set (ax(2), "yaxislocation", "right");
!   set (ax(2), "ycolor", get (h2(1), "color"));
    set (ax(2), "position", get (ax(1), "outerposition"));
    set (ax(2), "xlim", xlim);
  endfunction
--- 106,125 ----
  
    h2 = feval (fun2, x2, y2);
    set (ax(2), "yaxislocation", "right");
!   set (ax(2), "ycolor", getcolor (h2(1)));
    set (ax(2), "position", get (ax(1), "outerposition"));
    set (ax(2), "xlim", xlim);
  endfunction
+ 
+ function color = getcolor (ax)
+   obj = get (ax);
+   if (isfield (obj, "color"))
+     color = obj.color;
+   elseif (isfield (obj, "facecolor") && ! ischar (obj.facecolor))
+     color = obj.facecolor;
+   elseif (isfield (obj, "edgecolor") && !  ischar (obj.edgecolor))
+     color = obj.edgecolor;
+   else
+     color = [0, 0, 0];
+   endif
+ endfunction
Index: scripts/plot/Makefile.in
===================================================================
RCS file: /usr/local/cvsroot/octave/scripts/plot/Makefile.in,v
retrieving revision 1.58
diff -c -r1.58 Makefile.in
*** scripts/plot/Makefile.in    29 Nov 2007 20:50:24 -0000      1.58
--- scripts/plot/Makefile.in    29 Nov 2007 23:00:11 -0000
***************
*** 108,113 ****
--- 108,114 ----
    ndgrid.m \
    newplot.m \
    orient.m \
+   pareto.m \
    patch.m \
    pcolor.m \
    peaks.m \
*** scripts/plot/pareto.m.orig35        2007-11-29 23:58:25.434874621 +0100
--- scripts/plot/pareto.m       2007-11-29 23:48:31.593672162 +0100
***************
*** 0 ****
--- 1,117 ----
+ ## Copyright (C) 2007 David Bateman
+ ## Copyright (C) 2003 Alberto Terruzzi
+ ##
+ ## This file is part of Octave.
+ ##
+ ## Octave is free software; you can redistribute it and/or modify it
+ ## under the terms of the GNU General Public License as published by
+ ## the Free Software Foundation; either version 3 of the License, or (at
+ ## your option) any later version.
+ ##
+ ## Octave is distributed in the hope that it will be useful, but
+ ## WITHOUT ANY WARRANTY; without even the implied warranty of
+ ## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ ## General Public License for more details.
+ ##
+ ## You should have received a copy of the GNU General Public License
+ ## along with Octave; see the file COPYING.  If not, see
+ ## <http://www.gnu.org/licenses/>.
+ 
+ ## -*- texinfo -*-
+ ## @deftypefn {Function File} {} pareto (@var{x})
+ ## @deftypefnx {Function File} {} pareto (@var{x}, @var{y})
+ ## @deftypefnx {Function File} {} pareto (@var{h}, @dots{})
+ ## @deftypefnx {Function File} address@hidden =} pareto (@dots{})
+ ## Draw a Pareto chart, also called ABC chart. A Pareto chart is a bar graph 
+ ## used to arrange information in such a way that priorities for process 
+ ## improvement can be established. It organizes and displays information 
+ ## to show the relative importance of data. The chart is similar to the 
+ ## histogram or bar chart, except that the bars are arranged in decreasing 
+ ## order from left to right along the abscissa.
+ ## 
+ ## The fundamental idea (Pareto principle) behind the use of Pareto 
+ ## diagrams is that the majority of an effect is due to a small subset of the
+ ## causes, so for quality improvement the first few (as presented on the 
+ ## diagram) contributing causes to a problem usually account for the majority 
+ ## of the result. Thus, targeting these "major causes" for elimination 
+ ## results in the most cost-effective improvement scheme.
+ ##
+ ## The data are passed as @var{x} and the abscissa as @var{y}. If @var{y} is
+ ## absent, then the abscissa are assumed to be @code{1 : length (@var{x})}.
+ ## @var{y} can be a string array, a cell array of strings or a numerical
+ ## vector.
+ ##
+ ## An example of the use of @code{pareto} is
+ ##
+ ## @example
+ ## @group
+ ## Cheese = @{"Cheddar", "Swiss", "Camembert", "Munster", "Stilton", ...
+ ##    "Blue"@};
+ ## Sold = [105, 30, 70, 10, 15, 20];
+ ## pareto(Sold, Cheese);
+ ## @end group
+ ## @end example
+ ## @end deftypefn
+ 
+ function h = pareto (varargin)
+ 
+   [ax, varargin, nargin] = __plt_get_axis_arg__ ("pareto", varargin{:});
+ 
+   if (nargin != 1 && nargin != 2)
+     print_usage ();
+   endif
+ 
+   x = varargin {1}(:).';
+   if (nargin == 2)
+     y = varargin {2}(:).';
+     if (! iscell (y))
+       if (ischar (y))
+       y = cellstr (y);
+       else
+       y = num2cell (y);
+       endif
+     endif
+   else
+     y = cellfun (@(x) int2str (x), num2cell (1 : numel(x)), 
+                "UniformOutput", false);
+   endif
+ 
+   [x, idx] = sort (x, "descend");
+   y = y (idx);
+   cdf = cumsum (x);
+   maxcdf = cdf(end);
+   cdf = cdf ./ cdf (end);
+   [dummy, idx95] = min (abs (cdf - .95));
+   idx95 - idx95(1);
+ 
+   [ax, hbar, hline] = plotyy (ax, 1 : idx95, x (1 : idx95), 
+                             1 : length(cdf), 100 .* cdf, 
+                             @bar, @plot);
+ 
+   axis (ax(1), [1 - 0.6, idx95 + 0.6, 0, maxcdf]);
+   axis (ax(2), [1 - 0.6, idx95 + 0.6, 0, 100]);
+   set (ax(2), "ytick", [0, 20, 40, 60, 80, 100], 
+        "yticklabel", {"0%", "20%", "40%", "60%", "80%", "100%"});
+   set (ax(1), "xtick", 1 : idx95, "xticklabel", y (1: idx95));
+   set (ax(2), "xtick", 1 : idx95, "xticklabel", y (1: idx95));
+ 
+   if (nargout > 0)
+     h = [hbar; hline];
+   endif
+   
+ endfunction
+ 
+ %!demo
+ %! close
+ %! Cheese = {"Cheddar", "Swiss", "Camembert", "Munster", "Stilton", "Blue"};
+ %! Sold = [105, 30, 70, 10, 15, 20];
+ %! pareto(Sold, Cheese);
+ 
+ %!demo
+ %! close
+ %! % Suppose that we want establish which products makes 80 % of turnover.
+ %! Codes = {"AB4","BD7","CF8","CC5","AD11","BB5","BB3","AD8","DF3","DE7"};
+ %! Value = [2.35 7.9 2.45 1.1 0.15 13.45 5.4 2.05 0.85  1.65]';
+ %! SoldUnits = [54723 41114 16939 1576091 168000 687197 120222 168195, ...
+ %!              1084118 55576]';
+ %! pareto (Value.*SoldUnits, Codes);

2007-11-29  David Bateman  <address@hidden>

        * plot/pareto.m: New file.
        * plot/Makefile.in (SOURCES): Add it to the sources.
        * plot/__go_draw_axes__.m (do_tics1): Replace "%" with "%%" in tic
        marks to avoid gnuplot error about formating. More colorspec to
        after the tics.
        * plot/plotyy.m: More generic check for appropriate axis color.