Whisker plot/boxplot, how to?

Question

I made the graphics (code below), but my revisors ask for boxplot graphic. I confess I have no idea how to do it.

My database

Author min avg max mediana deviation
ClercA 100.1900 131.5194 180.2100 130.2000 16.0537
ClercB 100.1900 131.9290 190.2000 130.2000 15.6576
ClercC 100.1900 145.2007 210.2000 140.2100 17.3981
ClercD 100.1900 113.8061 150.2100 110.2000 10.8342
ClercE 100.1900 131.7290 180.2000 130.2000 15.7357
EberhartA 100.1900 129.8890 170.2100 130.2000 15.8132
EberhartB 100.1900 127.4685 170.2000 130.1900 14.4167
EberhartC 100.2000 142.6508 190.2100 140.2000 16.6241
EberhartD 100.1800 113.9665 150.2100 110.2000 11.2164
EberhartE 100.1900 129.6588 190.2100 130.2000 15.0647
XinA 100.1900 126.8997 170.2000 130.2000 14.3637
XinB 100.1900 124.3386 170.2000 130.1900 13.5221
XinC 100.1900 137.3308 190.2000 130.2100 16.0478
XinD 100.1800 113.8771 160.2000 110.2000 11.1794
XinE 100.1800 126.7993 180.2100 130.1900 14.4938

My latex code

\documentclass{article}
\usepackage{pgfplots}
\pgfplotsset{compat=newest}
\usepackage{adjustbox}
\pagestyle{empty}
\usetikzlibrary{positioning, backgrounds}

\begin{document}


    \begin{figure}[h!]%{width=\textwidth}
    \begin{adjustbox}{width=\textwidth}
        \centering
        \rotatebox{90}{%
            \begin{tikzpicture}[framed]
            \begin{axis}[
            xbar stacked, 
            xlabel={fitness},
            legend style={at={(0.5,-0.05)},anchor=north,legend columns=-1},
            symbolic y coords={ClercA,ClercB,ClercC,ClercD,ClercE,
                EberhartA, EberhartB, EberhartC, EberhartD, EberhartE,
                XinA, XinB, XinC, XinD, XinE},
            ytick=data,
            %enlarge y limits={abs=.70cm},
            %bar width={0.15cm},
            y=.45cm,
            y tick label style={rotate=-45,anchor=east,scale=0.5,font=\bfseries},
            %yticklabel style={\textbf{\pgfplotsretval},
            %every node near coord/.append style={anchor=south},
            nodes near coords, nodes near coords align={vertical},
           nodes near coords style={scale=0.35,font=\bfseries,text=black,rotate=-90}]


            %% mínimo
            \addplot+[xbar, violet!40] coordinates 
            {(100.1900,ClercA) (100.1900,ClercB) (100.1900,ClercC) (100.1800,ClercD)(100.1900,ClercE)

                (100.1900,EberhartA) (100.1900,EberhartB) (100.2000,EberhartC) (100.1800,EberhartD) (100.1900,EberhartE)

                (100.1900,XinA) (100.1900,XinB) (100.1900,XinC) (100.1800,XinD) (100.1800,XinE) }; %min

            %% average
            \addplot+[xbar, red!70] coordinates 
            {(131.5194,ClercA) (131.9290,ClercB) (145.2007,ClercC) (113.8061,ClercD) (131.7290,ClercE)

                (129.8890,EberhartA) (127.4685,EberhartB) (142.6508,EberhartC) (113.9665,EberhartD) (129.6588,EberhartE)

                (126.8997,XinA) (124.3386,XinB) (137.3308,XinC) (113.8771,XinD) (126.7993,XinE) }; %avg

            %% máximo
            \addplot+[xbar, green!50] coordinates 
            {(180.2100,ClercA) (190.2000,ClercB) (210.2000,ClercC) (150.2100,ClercD) (180.2000,ClercE)

                (170.2100,EberhartA) (170.2000,EberhartB) (190.2100,EberhartC) (150.2100,EberhartD) (190.2100,EberhartE)

                (170.2000,XinA) (170.2000,XinB) (190.2000,XinC) (160.2000,XinD) (180.2100,XinE) }; %max

            %% standart deviation
            \addplot+[xbar,nodes near coords, nodes near coords align={center}, purple!40] coordinates 
            {(16.0537,ClercA) (15.6576,ClercB) (17.3981,ClercC) (10.8342,ClercD)(15.7357,ClercE)

                (15.8132,EberhartA) (14.4167,EberhartB) (16.6241,EberhartC) (11.2164,EberhartD) (15.0647,EberhartE)

                (14.3637,XinA) (13.5221,XinB) (16.0478,XinC) (11.1794,XinD) (14.4938,XinE) }; %deviation

            %%mediana
            \addplot+[xbar, blue!40] coordinates 
            {(130.2000,ClercA) (130.2000,ClercB) (140.2100,ClercC) (110.2000,ClercD)(130.2000,ClercE)

                (130.2000,EberhartA) (130.1900,EberhartB) (140.2000,EberhartC) (110.2000,EberhartD) (130.2000,EberhartE)

                (130.2000,XinA) (130.1900,XinB) (130.2100,XinC) (110.2000,XinD) (130.1900,XinE) }; %mediana



            \legend{min,avg,max, deviation, mediana}    

            \end{axis}
            \end{tikzpicture}}
    \end{adjustbox}
    \caption{Scenario settings: a) $c_1 = 1$ and $c_2 = 2$; b) $c_1 = 2$ and $c_2 = 1$;
        c) $c_l = c_2 = 1$; d) $c_1 = c_2 = 2$ and e) $c_1 = c_2 = 1.496$.}
    \label{fig:simulation_results_fitness_gpop100}
    %\end{subfigure}%
\end{figure}


\begin{figure}[h!]
    \begin{adjustbox}{width=\textwidth}

            \begin{tikzpicture}[framed]
            \begin{axis}[
            xbar stacked, 
            xlabel={fitness},
            legend style={at={(0.5,-0.05)},anchor=north,legend columns=-1},
            symbolic y coords={ClercA,ClercB,ClercC,ClercD,ClercE,
                EberhartA, EberhartB, EberhartC, EberhartD, EberhartE,
                XinA, XinB, XinC, XinD, XinE},
            ytick=data,
            %enlarge y limits={abs=.70cm},
            %bar width={0.15cm},
            y=.45cm,
            y tick label style={anchor=east,scale=0.5,font=\bfseries},
            %every node near coord/.append style={anchor=center},
            nodes near coords,  nodes near coords style={scale=0.35,text=black,rotate=-90},]


            %% mínimo
            \addplot+[xbar, nodes near coords style={anchor=south},violet!40] coordinates 
            {(100.1900,ClercA) (100.1900,ClercB) (100.1900,ClercC) (100.1800,ClercD)(100.1900,ClercE)

                (100.1900,EberhartA) (100.1900,EberhartB) (100.2000,EberhartC) (100.1800,EberhartD) (100.1900,EberhartE)

                (100.1900,XinA) (100.1900,XinB) (100.1900,XinC) (100.1800,XinD) (100.1800,XinE) } ; %min

            %% average
            \addplot+[xbar,red!70] coordinates 
            {(131.5194,ClercA) (131.9290,ClercB) (145.2007,ClercC) (113.8061,ClercD) (131.7290,ClercE)

                (129.8890,EberhartA) (127.4685,EberhartB) (142.6508,EberhartC) (113.9665,EberhartD) (129.6588,EberhartE)

                (126.8997,XinA) (124.3386,XinB) (137.3308,XinC) (113.8771,XinD) (126.7993,XinE) }; %avg

            %% máximo
            \addplot+[xbar,green!50] coordinates 
            {(180.2100,ClercA) (190.2000,ClercB) (210.2000,ClercC) (150.2100,ClercD) (180.2000,ClercE)

                (170.2100,EberhartA) (170.2000,EberhartB) (190.2100,EberhartC) (150.2100,EberhartD) (190.2100,EberhartE)

                (170.2000,XinA) (170.2000,XinB) (190.2000,XinC) (160.2000,XinD) (180.2100,XinE) }; %max

            %% standart deviation
            \addplot+[xbar,purple!40] coordinates 
            {(16.0537,ClercA) (15.6576,ClercB) (17.3981,ClercC) (10.8342,ClercD)(15.7357,ClercE)

                (15.8132,EberhartA) (14.4167,EberhartB) (16.6241,EberhartC) (11.2164,EberhartD) (15.0647,EberhartE)

                (14.3637,XinA) (13.5221,XinB) (16.0478,XinC) (11.1794,XinD) (14.4938,XinE) }; %deviation

            %%mediana
            \addplot+[xbar,blue!40] coordinates 
            {(130.2000,ClercA) (130.2000,ClercB) (140.2100,ClercC) (110.2000,ClercD)(130.2000,ClercE)

                (130.2000,EberhartA) (130.1900,EberhartB) (140.2000,EberhartC) (110.2000,EberhartD) (130.2000,EberhartE)

                (130.2000,XinA) (130.1900,XinB) (130.2100,XinC) (110.2000,XinD) (130.1900,XinE) }; %mediana



            \legend{min,avg,max, deviation, mediana}    

            \end{axis}
            \end{tikzpicture}
    \end{adjustbox}
    \caption{Scenario settings: a) $c_1 = 1$ and $c_2 = 2$; b) $c_1 = 2$ and $c_2 = 1$;
        c) $c_l = c_2 = 1$; d) $c_1 = c_2 = 2$ and e) $c_1 = c_2 = 1.496$.}
    \label{fig:simulation_results_fitness_gpop100}
    %\end{subfigure}%
\end{figure}

\end{document}

score 1 · Answer 1 · 2019-06-10T20:59:44.313

This is a way to convert your table to some boxplot. (Unfortunately I was unable to apply this great answer to your case.) It does not compete with the great look of your own plot.

\documentclass[tikz,border=3.14mm]{standalone}
\usepackage{pgfplotstable}
\pgfplotsset{compat=1.16}
\usepgfplotslibrary{statistics}
\pgfplotstableread{%
Author min avg max mediana deviation
ClercA 100.1900 131.5194 180.2100 130.2000 16.0537
ClercB 100.1900 131.9290 190.2000 130.2000 15.6576
ClercC 100.1900 145.2007 210.2000 140.2100 17.3981
ClercD 100.1900 113.8061 150.2100 110.2000 10.8342
ClercE 100.1900 131.7290 180.2000 130.2000 15.7357
EberhartA 100.1900 129.8890 170.2100 130.2000 15.8132
EberhartB 100.1900 127.4685 170.2000 130.1900 14.4167
EberhartC 100.2000 142.6508 190.2100 140.2000 16.6241
EberhartD 100.1800 113.9665 150.2100 110.2000 11.2164
EberhartE 100.1900 129.6588 190.2100 130.2000 15.0647
XinA 100.1900 126.8997 170.2000 130.2000 14.3637
XinB 100.1900 124.3386 170.2000 130.1900 13.5221
XinC 100.1900 137.3308 190.2000 130.2100 16.0478
XinD 100.1800 113.8771 160.2000 110.2000 11.1794
XinE 100.1800 126.7993 180.2100 130.1900 14.4938
}\datatable

\begin{document}
\begin{tikzpicture}
\pgfplotstablegetrowsof{\datatable}
\pgfmathtruncatemacro{\rownumber}{\pgfplotsretval-1}
\begin{axis}[boxplot/draw direction=x,yticklabels={ClercA,ClercB,ClercC,ClercD,ClercE,
                EberhartA, EberhartB, EberhartC, EberhartD, EberhartE,
                XinA, XinB, XinC, XinD, XinE},ytick={1,...,\the\numexpr\rownumber+1},height=12cm
                ]
\typeout{\rownumber}
\pgfplotsinvokeforeach{0,...,\rownumber}{
 \pgfplotstablegetelem{#1}{min}\of\datatable
 \edef\mymin{\pgfplotsretval}
 \pgfplotstablegetelem{#1}{avg}\of\datatable
 \edef\myavg{\pgfplotsretval}
 \pgfplotstablegetelem{#1}{max}\of\datatable
 \edef\mymax{\pgfplotsretval}
 \pgfplotstablegetelem{#1}{mediana}\of\datatable
 \edef\mymediana{\pgfplotsretval}
 \pgfplotstablegetelem{#1}{deviation}\of\datatable
 \edef\mydeviation{\pgfplotsretval}
 \typeout{\mymin,\mymax,\myavg,\mymediana,\mydeviation}
 \pgfmathsetmacro{\mylowerq}{\mymediana-\mydeviation}
 \pgfmathsetmacro{\myupperq}{\mymediana+\mydeviation}
 \edef\temp{\noexpand\addplot+[
    boxplot prepared={
     lower whisker=\mymin,
     upper whisker=\mymax,
     lower quartile=\mylowerq,
     upper quartile=\myupperq,
     median=\myavg
  }
  ]coordinates {};}
 \temp
}
\end{axis}
\end{tikzpicture}
\end{document}

tanks for help! I only correct foreach to start in 0. And you exemple is better... I understood your answer and how this type of chart works. Next time I know how to do it! — Marcos Laureano, Jun 10 '19 at 20:49

score 0 · Answer 2 · answered Jun 10 '19 at 23:52

after a little play, I got results that were exactly what my supervisor wanted! @marmot, tanks again!

\documentclass[crop=false]{standalone}
\usepackage{pgfplotstable}
\usepgfplotslibrary{colorbrewer}
%\pgfplotsset{compat=1.16}
\usepgfplotslibrary{statistics}
\pgfplotstableread{%
Author min avg max mediana deviation
ClercA 100.1900 131.5194 180.2100 130.2000 16.0537
ClercB 100.1900 131.9290 190.2000 130.2000 15.6576
ClercC 100.1900 145.2007 210.2000 140.2100 17.3981
ClercD 100.1900 113.8061 150.2100 110.2000 10.8342
ClercE 100.1900 131.7290 180.2000 130.2000 15.7357
EberhartA 100.1900 129.8890 170.2100 130.2000 15.8132
EberhartB 100.1900 127.4685 170.2000 130.1900 14.4167
EberhartC 100.2000 142.6508 190.2100 140.2000 16.6241
EberhartD 100.1800 113.9665 150.2100 110.2000 11.2164
EberhartE 100.1900 129.6588 190.2100 130.2000 15.0647
XinA 100.1900 126.8997 170.2000 130.2000 14.3637
XinB 100.1900 124.3386 170.2000 130.1900 13.5221
XinC 100.1900 137.3308 190.2000 130.2100 16.0478
XinD 100.1800 113.8771 160.2000 110.2000 11.1794
XinE 100.1800 126.7993 180.2100 130.1900 14.4938
}\datatable

\begin{document}
\begin{tikzpicture}
\begin{axis}[boxplot/draw direction=y,
xticklabels={ClercA,ClercB,ClercC,ClercD,ClercE,EberhartA, EberhartB, EberhartC, EberhartD, EberhartE,XinA, XinB, XinC, XinD, XinE},
xtick={1,2,3,4,5,6,7,8,9,10,11,12,13,14,15},
x tick label style={scale=0.5,font=\bfseries, rotate=60,,align=center},
ylabel={fitness},cycle list/YlGnBu-5 ]


\pgfplotstablegetrowsof{\datatable}
\pgfmathtruncatemacro{\rownumber}{\pgfplotsretval-1}
\pgfplotsinvokeforeach{0,...,\rownumber}{

    \pgfplotstablegetelem{#1}{min}\of\datatable
    \edef\mymin{\pgfplotsretval}

    \pgfplotstablegetelem{#1}{avg}\of\datatable
    \edef\myavg{\pgfplotsretval}

    \pgfplotstablegetelem{#1}{max}\of\datatable
    \edef\mymax{\pgfplotsretval}

    \pgfplotstablegetelem{#1}{mediana}\of\datatable
    \edef\mymediana{\pgfplotsretval}

    \pgfplotstablegetelem{#1}{deviation}\of\datatable
    \edef\mydeviation{\pgfplotsretval}

    \typeout{\mymin,\mymax,\myavg,\mymediana,\mydeviation}
    \pgfmathsetmacro{\mylowerq}{\mymediana-\mydeviation}
    \pgfmathsetmacro{\myupperq}{\mymediana+\mydeviation}
    \edef\temp{\noexpand\addplot+[,
        boxplot prepared={
            lower whisker=\mymin,
            upper whisker=\mymax,
            lower quartile=\mylowerq,
            upper quartile=\myupperq,
            median=\myavg,
            every box/.style={solid,fill,opacity=0.5},
            every whisker/.style={solid },
            every median/.style={solid},
        }, 
        ]coordinates {};}
    \temp
}
\end{axis}
\end{tikzpicture}

\begin{tikzpicture}
\pgfplotstablegetrowsof{\datatable}
\pgfmathtruncatemacro{\rownumber}{\pgfplotsretval-1}
\begin{axis}[boxplot/draw direction=y,
xticklabels={ClercA,ClercB,ClercC,ClercD,ClercE,EberhartA, EberhartB, EberhartC, EberhartD, EberhartE,XinA, XinB, XinC, XinD, XinE},
xtick={1,...,\the\numexpr\rownumber+1},
x tick label style={scale=0.5,font=\bfseries, rotate=60,align=center},
ylabel={fitness},cycle list/Set1-5 ]


\pgfplotsinvokeforeach{0,...,\rownumber}{

    \pgfplotstablegetelem{#1}{min}\of\datatable
    \edef\mymin{\pgfplotsretval}

    \pgfplotstablegetelem{#1}{avg}\of\datatable
    \edef\myavg{\pgfplotsretval}

    \pgfplotstablegetelem{#1}{max}\of\datatable
    \edef\mymax{\pgfplotsretval}

    \pgfplotstablegetelem{#1}{mediana}\of\datatable
    \edef\mymediana{\pgfplotsretval}

    \pgfplotstablegetelem{#1}{deviation}\of\datatable
    \edef\mydeviation{\pgfplotsretval}

    \typeout{\mymin,\mymax,\myavg,\mymediana,\mydeviation}
    \pgfmathsetmacro{\mylowerq}{\mymediana-\mydeviation}
    \pgfmathsetmacro{\myupperq}{\mymediana+\mydeviation}
    \edef\temp{\noexpand\addplot+[,
        boxplot prepared={
            lower whisker=\mymin,
            upper whisker=\mymax,
            lower quartile=\mylowerq,
            upper quartile=\myupperq,
            median=\myavg,
            every box/.style={solid,fill,opacity=0.5},
            every whisker/.style={solid },
            every median/.style={solid},
        }, 
        ]coordinates {};}
    \temp
}
\end{axis}
\end{tikzpicture}
\end{document}

Whisker plot/boxplot, how to?

2 Answers2