[Boost-commit] svn:boost r50519 - sandbox/SOC/2007/visualization/libs/svg_plot/example

Subject: [Boost-commit] svn:boost r50519 - sandbox/SOC/2007/visualization/libs/svg_plot/example
From: pbristow_at_[hidden]
Date: 2009-01-08 12:06:33

Author: pbristow
Date: 2009-01-08 12:06:32 EST (Thu, 08 Jan 2009)
New Revision: 50519
URL: http://svn.boost.org/trac/boost/changeset/50519

Log:
1st commit
Added:
   sandbox/SOC/2007/visualization/libs/svg_plot/example/demo_Hoaglin.cpp   (contents, props changed)
   sandbox/SOC/2007/visualization/libs/svg_plot/example/demo_boxplot.cpp   (contents, props changed)

Added: sandbox/SOC/2007/visualization/libs/svg_plot/example/demo_Hoaglin.cpp
==============================================================================
--- (empty file)
+++ sandbox/SOC/2007/visualization/libs/svg_plot/example/demo_Hoaglin.cpp	2009-01-08 12:06:32 EST (Thu, 08 Jan 2009)
@@ -0,0 +1,266 @@
+// demo_boxplot_full.cpp
+
+// Copyright Jacob Voytko 2007
+// Copyright Paul A. Bristow 2008
+
+// Use, modification and distribution are subject to the
+// Boost Software License, Version 1.0.
+// (See accompanying file LICENSE_1_0.txt
+// or copy at http://www.boost.org/LICENSE_1_0.txt)
+
+// An example to demonstrate the implementation of boxplot.
+
+// This file is written to be included from a Quickbook .qbk document.
+// It can be compiled by the C++ compiler, and run. Any output can
+// also be added here as comment or included or pasted in elsewhere.
+
+// Caution: this file contains Quickbook markup as well as code
+// and comments: don't change any of the special comment markups!
+
+//[demo_Hoaglin_1
+
+/*`
+
+"Some Implementations of the Boxplot"
+Michael Frigge, David C. Hoaglin and Boris Iglewicz
+The American Statistician, Vol. 43, No. 1 (Feb., 1989), pp. 50-54
+discusses the design of the boxplot.
+
+However the plot of their example data shown below shows the considerable variation in the appearance of the same data,
+using different definitions of quartiles used in various popular statistics packages.
+
+One obvious conclusion is that you should not expect boxplots to look the same when using more than one program.
+
+Boost.Plot provides 5 popular definitions for the quartiles.
+This should allow the user to produce plots that look similar to boxplots from most statistics plotting program.
+To confuse matter further, most have their own default definition *and* options to chose other definitions:
+these options are shown below as type, method, PCTLDEF.
+
+The interquartile range is calculated using the 1st & 3rd sample quartiles,
+but there are various ways to calculate those quartiles, summarised in
+Rob J. Hyndman and Yanan Fan, 1996, "Sample Quantiles in Statistical Packages",
+The American Statistician 50(4):361-365, (1996).
+
+The interquartile range, often called IQR is quartile 3 (p = 3/4) - quartile 1 (1/4).
+The median is the 2nd quartile (p = 2/4 = 1/2).
+
+Five of Hyndman and Fan's sample quantile definitions have a particularly simple common form
+selected according to which definition of m is chosen in function quantiles.
+This is implemented in function quantiles by parameter `HF_definition`:
+
+  double quantile(vector<double>& data, double p, int HF_definition = 8);
+
+The default definition is that recommended by Hyndman and Fan, or
+users can select which definition is used for all boxplots, or individual data series as shown in the example below.
+
+   my_boxplot.quartile_definition(5); // All plots 
+
+   my_boxplot.plot.quartile_definition(7); // Just this data series plot.
+
+Hyndman and Fan definitions 4 to 8 are used by the following packages:
+
+* #4 SAS (PCTLDEF=1), R (type=4), Maple (method=3)
+* #5 R (type=5), Maple (method=4), Wolfram Mathematica quartiles.
+* #6 Minitab, SPSS, BMDP, JMP, SAS (PCTLDEF=4), R(type=6), Maple (method=5).
+* #7 Excel, S-Plus, R (type=7[default]), Maxima, Maple (method=6).
+* #8 H&F 8: R (type=8), Maple (method=7[default]).
+
+Some observations on the various options are:
+
+* #4 Often a moderate interquartile range.
+
+* #5 Symmetric linear interpolation: a common choice when the data represent a sample
+from a continuous distribution and you want an unbiased estimate of the quartiles of that distribution.
+
+* #6 This "half" sample excludes the sample median (k observations) for odd n (=2*k+1).
+This will tend to be a better estimate for the population quartiles,
+but will tend to give quartile estimates that are a bit too far
+from the center of the whole sample (too wide an interquartile range).
+
+* #7 Smallest interquartile range, so flags most outliers.
+For a continuous distribution,
+this will tend to give too narrow an interquartile range,
+since there will tend to be a small fraction of the population beyond the extreme
+sample observations. In particular, for odd n (=2*k+1), Excel calculates the
+1st (3rd) quartile as the median of the lower (upper) "half" of the sample
+including the sample median (k+1 observations).
+
+* #8 recommended by H&F because it is
+approximately median-unbaised estimate regardless of distribution
+and thus suitable for continuous and discrete distributions.
+which gives quartiles between those reported by Minitab and Excel.
+This approach is approximately median unbiased for continuous distributions.
+Slightly higher interquartile range than definition 7.
+
+The 'fences' beyond which points are regarded as outliers, or extreme outliers,
+are a multiplying factor, usually called k, and usually 1.5 * interquartile range,
+and 3 * interquartile range as recommended by Hoaglin et al.
+
+*/
+
+#include <vector>
+using std::vector;
+#include <cmath>
+using ::sin;
+//#include <boost/assert.hpp> // for BOOST_ASSERT
+#include <boost/svg_plot/svg_boxplot.hpp>
+
+#include <boost/svg_plot/quantile.hpp>
+using boost::svg::quantile;
+
+// double boost::svg::quantile(vector<double>& data, double p, int HF_definition);
+// Estimate pth quantile of data using one of 5 definitions.
+// Default HF_definition is the recommendation of Hyndman and Fan, definition #8.
+
+#include <boost/array.hpp>
+  using boost::array;
+
+#include <iostream>
+using std::cout;
+using std::endl;
+
+//] [demo_Hoaglin_1]
+
+int main()
+{
+  using namespace boost::svg;
+  try
+  {
+//[demo_Hoaglin_2]
+  // 11 values from Hoaglin et al page 50.
+  const boost::array<double, 11> Hoaglin_data = {53., 56., 75., 81., 82., 85., 87., 89., 95., 99., 100.};
+  //                                                       q1           median           q3
+
+  vector<double> Hoaglin(Hoaglin_data.begin(), Hoaglin_data.end()); 
+  for (int def = 4; def <= 8; def++)
+  { // All the F&Y definitions of quartiles.
+    double q1 = quantile(Hoaglin, 0.25, def); // 75
+    double q2 = quantile(Hoaglin, 0.5, def); // 85
+    double q3 = quantile(Hoaglin, 0.75, def); // 95
+    cout << "Hoaglin definition #" << def << ", q1 " << q1 
+      << ", q2 " << q2 << ", q3 " << q3 << ", IQR " << q3 - q1 << endl;
+  } // for
+
+  // Same data copied for different data series.
+  vector<double> Hoaglin4(Hoaglin_data.begin(), Hoaglin_data.end()); 
+  vector<double> Hoaglin5(Hoaglin_data.begin(), Hoaglin_data.end()); 
+  vector<double> Hoaglin6(Hoaglin_data.begin(), Hoaglin_data.end()); 
+  vector<double> Hoaglin7(Hoaglin_data.begin(), Hoaglin_data.end()); 
+  vector<double> Hoaglin8(Hoaglin_data.begin(), Hoaglin_data.end()); 
+
+  svg_boxplot H_boxplot;
+
+  /*`Show the quartile definition default.
+*/
+    cout << "Default boxplot.quartile_definition() = " << H_boxplot.quartile_definition() << endl; // 8
+
+/*` Add title, labels, range etc to the whole boxplot:
+*/
+  H_boxplot  // Title and axes labels.
+    .title("Hoaglin Example Data")
+    .x_label("Boxplot")
+    .y_label("Value")
+    .y_range(45, 115)  // Y-Axis range.
+    .y_minor_tick_length(2)
+    .y_major_interval(10);
+
+/*`Add a few setting to the plot including setting quartile definition (though is actually same as the default 8),
+and show that the value is stored.
+*/ 
+    svg_boxplot& b = H_boxplot.median_values_on(true)
+    .outlier_values_on(true)
+    .extreme_outlier_values_on(true)
+    .quartile_definition(8)
+    ;
+/*`Show the quartile definition just assigned:
+*/
+    cout << "boxplot.quartile_definition() = " << b.quartile_definition() << endl; // 8
+
+/*`Add a data series container, and labels, to the plot using the whole boxplot quartile definition set.
+*/
+    H_boxplot.plot(Hoaglin_data, "default_8");
+
+/*`Add another data series container, and the labels, to the plot, and select a *different* quartile definition.
+*/
+
+    svg_boxplot_series& d4 =
+    H_boxplot.plot(Hoaglin4, "def #4")
+    .whisker_length(4.)
+    .quartile_definition(4);
+
+/*`Show the quartile definition just assigned to the this data series.
+*/
+  cout << "boxplot_series.quartile_definition() = " << d4.quartile_definition() << endl; // 4
+
+/*`Add yet more data series container, and the labels, to the plot, and select a *different* quartile definition for each.
+*/    H_boxplot.plot(Hoaglin5, "def #5")
+    .whisker_length(5.)
+    .quartile_definition(5);
+
+    H_boxplot.plot(Hoaglin6, "def #6")
+    .whisker_length(6.)
+    .quartile_definition(6);
+
+    H_boxplot.plot(Hoaglin6, "def #7")
+    .whisker_length(7.)
+    .quartile_definition(7);
+
+    H_boxplot.plot(Hoaglin6, "def #8")
+    .whisker_length(8.)
+    .quartile_definition(8);
+
+/*`Write the entire SVG plot to a file.
+*/
+  H_boxplot.write("demo_Hoaglin.svg");
+
+/*`You can view the plot at [@demo_Hoaglin.svg].
+*/
+
+//] [demo_Hoaglin_2]
+  }
+  catch(const std::exception& e)
+  {
+    std::cout <<
+      "\n""Message from thrown exception was:\n  " << e.what() << std::endl;
+  }
+  return 0;
+} // int main()
+
+/*
+
+Output:
+
+//[demo_Hoaglin_output
+
+demo_Hoaglin.cpp
+Linking...
+Embedding manifest...
+Autorun "j:\Cpp\SVG\debug\demo_Hoaglin.exe"
+Hoaglin definition #4, q1 70.25, q2 83.5, q3 90.5, IQR 20.25
+Hoaglin definition #5, q1 76.5, q2 85, q3 93.5, IQR 17
+Hoaglin definition #6, q1 75, q2 85, q3 95, IQR 20
+Hoaglin definition #7, q1 78, q2 85, q3 92, IQR 14
+Hoaglin definition #8, q1 76, q2 85, q3 94, IQR 18
+Default boxplot.quartile_definition() = 8
+boxplot.quartile_definition() = 8
+boxplot_series.quartile_definition() = 4
+Min outlier fences <49 or >121
+Min extreme fences <22 or >148
+Min outlier fences <39.875 or >120.875
+Min extreme fences <9.5 or >151.25
+Min outlier fences <51 or >119
+Min extreme fences <25.5 or >144.5
+Min outlier fences <45 or >125
+Min extreme fences <15 or >155
+Min outlier fences <57 or >113
+Min extreme fences <36 or >134
+Min outlier fences <49 or >121
+Min extreme fences <22 or >148
+Build Time 0:02
+Build log was saved at "file://j:\Cpp\SVG\demo_Hoaglin\Debug\BuildLog.htm"
+
+//] [demo_Hoaglin_output]
+
+
+*/
+

Added: sandbox/SOC/2007/visualization/libs/svg_plot/example/demo_boxplot.cpp
==============================================================================
--- (empty file)
+++ sandbox/SOC/2007/visualization/libs/svg_plot/example/demo_boxplot.cpp	2009-01-08 12:06:32 EST (Thu, 08 Jan 2009)
@@ -0,0 +1,154 @@
+// demo_boxplot.cpp
+
+// Copyright Jacob Voytko 2007
+// Copyright Paul A. Bristow 2008
+
+// Use, modification and distribution are subject to the
+// Boost Software License, Version 1.0.
+// (See accompanying file LICENSE_1_0.txt
+// or copy at http://www.boost.org/LICENSE_1_0.txt)
+
+// An example to demonstrate some use of boxplot options.
+// See also boxplot_full.cpp for a even wider range of use.
+
+// This file is written to be included from a Quickbook .qbk document.
+// It can be compiled by the C++ compiler, and run. Any output can
+// also be added here as comment or included or pasted in elsewhere.
+
+// Caution: this file contains Quickbook markup as well as code
+// and comments: don't change any of the special comment markups!
+
+//[demo_boxplot_1
+
+/*`
+Boxplot is a  convenient way of graphically depicting groups of numerical data 
+through their five-number summaries.
+Show 1st quartile, median and 3rd quartile as a box,
+95% confidence interval as whiskers,
+and outliers and extreme outliers.
+
+See [@http://en.wikipedia.org/wiki/Boxplot boxplot] and
+
+Some Implementations of the Boxplot
+Michael Frigge, David C. Hoaglin and Boris Iglewicz
+The American Statistician, Vol. 43, No. 1 (Feb., 1989), pp. 50-54
+
+First we need a few includes to use Boost.Plot.
+*/
+
+#include <vector>
+using std::vector;
+#include <cmath>
+using ::sin;
+#include <boost/svg_plot/svg_boxplot.hpp>
+
+#include <iostream>
+using std::cout;
+using std::endl;
+
+/*`Use two functions, 1/x and sin(x), to simulate distributions.
+*/
+
+double f(double x)
+{ // Effectively 1/x.
+  return 50 / x;
+}
+
+double g(double x)
+{ // Effectively sin(x).
+  return 40 + 25 * sin(x * 50);
+}
+//] [demo_boxplot_1]
+
+int main()
+{
+  using namespace boost::svg;
+  try
+  {
+//[demo_boxplot_2]
+/*`10 values are computed and stored in two std:: vectors.
+*/
+  std::vector<double> data1;
+  std::vector<double> data2;
+
+  cout.precision(2);
+  for(double i = 0.1; i < 10; i += 0.1)
+  {   // Fill our vectors with 100 values:
+    double fv = f(i);
+    double gv = g(i);
+    // cout << i << ' ' << fv << ' ' << gv << endl;
+    data1.push_back(fv);
+    data2.push_back(gv);
+  }
+
+/*`A new boxplot is contructed and several settings added.
+*/
+  svg_boxplot my_boxplot;
+
+  my_boxplot.background_border_color(darkblue);
+  my_boxplot.background_color(azure);
+
+  my_boxplot  // Title and axes labels.
+    .title("Boxplots of 1/x and sin(x) Functions")
+    .x_label("Functions")
+    .y_label("Population Size");
+
+  my_boxplot.y_range(0, 100)  // Y-Axis information.
+    .y_minor_tick_length(10)
+    .y_major_interval(20);
+
+   // box'n'whiskers options apply to the plot, AND for each data series boxplot
+   // so can be set for all boxplots or separately for each plot.
+   my_boxplot.plot(data1, "test").box_width(10).whisker_length(5).box_style().fill_color(pink).stroke_color(green);
+   // TODO these should be chainable like: .box_fill(pink), box_stroke(green)...
+   //my_boxplot.plot(data1, "test").box_width(10).whisker_length(5).median_style().stroke_color(purple);
+
+
+/*`Add the two data series containers, and their labels, to the plot.
+*/
+
+  my_boxplot.plot(data1, "[50 / x]");
+  my_boxplot.plot(data2, "[40+25*sin(x*50)]");
+
+/*  cout << "my_boxplot.title " << my_boxplot.title() << endl;
+  cout << "my_boxplot.x_label_text "<< my_boxplot.x_label_text() << endl;
+  cout << "my_boxplot.y_label_text " << my_boxplot.y_label_text() << endl; 
+
+  cout << "my_boxplot.background_color " << my_boxplot.background_color() << endl;
+  cout << "my_boxplot.background_border_color " << my_boxplot.background_border_color() << endl;
+  cout << "my_boxplot.plot_background_color " << my_boxplot.plot_background_color() << endl;
+ cout << "my_boxplot.plot_border_color " << my_boxplot.plot_border_color() << endl;
+*/ 
+
+/*`Finally write the SVG plot to a file.
+*/
+  my_boxplot.write("demo_boxplot.svg");
+/*`You can view the plot at demo_boxplot.svg."
+*/
+
+//] [demo_boxplot_2]
+  }
+  catch(const std::exception& e)
+  {
+    std::cout <<
+      "\n""Message from thrown exception was:\n  " << e.what() << std::endl;
+  }
+  return 0;
+} // int main()
+
+/*
+
+Output:
+
+
+Compiling...
+demo_boxplot.cpp
+Linking...
+Embedding manifest...
+Autorun "j:\Cpp\SVG\debug\demo_boxplot.exe"
+Build Time 0:02
+
+
+
+*/
+