[Tex/LaTex] Stacked bar plot, percentage of entire bar and the total value

pgfplotspgfplotstable

I have made quite a lot of progress of where I want to be, but there is still some fixes to be made. What I have right now as an image:

However, as we can see, the percentage are all relative to the Ref, which has 100%. Looking at the second bar (x-axis 4), we have 64.2+1.6+20.6 = 86.4 != 100. Additional to that, preferably the total time of all the stacks should show up above it. A rough sketch of what I'm thinking:

File data.txt:

one,two,three,four
Ref,0.0,194.228732386,0.0
4,40.0069652252,3.1391023845,124.752030532
8,31.416268219,4.0078951455,111.883188346
9,76.8558069667,3.9906615645,112.515372569
11,82.992369082,4.38474317025,112.11896845
13,81.5133020117,4.27987078675,116.598416893
16,81.5851874813,8.44496483425,117.332427258

Latex file:

\pgfplotstableread[col sep = comma]{data.txt}\mydata

\makeatletter

% To allow the 'Ref' x-axis entry
\pgfplotsset{
    /pgfplots/flexible xticklabels from table/.code n args={3}{%
        \pgfplotstableread[#3]{#1}\coordinate@table
        \pgfplotstablegetcolumn{#2}\of{\coordinate@table}\to\pgfplots@xticklabels
        \let\pgfplots@xticklabel=\pgfplots@user@ticklabel@list@x
    }
}

% For calculating the percentage
\pgfplotsset{
    compat=1.5.1,
    percentage ticks/.code={
        \pgfplotsset{
            xtick pos=left,
            axis x line*=left,
            xlabel=Units,
            nodes near coords={
                \pgfkeys{
                    /pgf/fpu=true,  % PGFplots uses floating points internally
                    /pgf/fpu/output format=fixed
                }
                \pgfmathparse{100*\pgfplotspointmeta/\pgfplots@axiswide@metamax}
                \pgfmathprintnumber{\pgfmathresult}\,\%
            },
            every node near coord/.append style={yshift=1ex}
        }
    }
}
\makeatother


\begin{tikzpicture}
\begin{axis}[
/pgf/number format/precision=1,
compat=newest, %Better label placement
ybar stacked,
width=1.2\textwidth,
height=0.4\textheight,
enlarge y limits={upper, value=0.2},
ymin=0,
enlarge x limits = 0.2,
bar width=32pt,
title={Instance els19 with 1\% max error},
legend style={at={(0.5,-0.15)},
anchor=north,legend columns=0},
ylabel={Time in milisseconds},
xtick=data,
flexible xticklabels from table={data.txt}{one}{col sep=comma},
xticklabel style={text height=1.5ex},
nodes near coords,
axis lines*=left,
cycle list={
    {fill=black!30,draw=black!30},
    {fill=black!20,draw=black!20},
    {fill=black!10,draw=black!10}
},
axis on top,
major grid style=white,
ymajorgrids,
legend style={draw=none,/tikz/every even column/.append style={column sep=0.5cm}},
percentage ticks
]
\addplot table[x expr=\coordindex, y=two]{\mydata};
\addplot table[x expr=\coordindex, y=three]{\mydata};
\addplot table[x expr=\coordindex, y=four]{\mydata};
\legend{Best 1\%,Average 1\%,Worst 1\%}
\end{axis}
\end{tikzpicture}

I have a feeling that the most interesting row is the:

\pgfmathparse{100*\pgfplotspointmeta/\pgfplots@axiswide@metamax}

However, I've looked through the manual for quite some time unable to find a suitable macro for the "current" bar.

Any help/pointers at all are appreciated!

Best Answer

As people have upvoted this, I figured I would respond with my solution:

I extended the data field with a sum, which was possible in my case. I believe you can also do this by creating a temporary table with \pgfplotstable as well.

% Showing the sum on top
\pgfplotsset{
        show sum on top/.style={
            /pgfplots/scatter/@post marker code/.append code={%
                \pgfkeys{
                    /pgf/fpu=true,  % PGFplots uses floating points internally
                    /pgf/fpu/output format=fixed
                }
                \pgfmathparse{\pgfkeysvalueof{/data point/x}}
                \pgfmathtruncatemacro\xidx{\pgfmathresult}
                \pgfplotstablegetelem{\xidx}{sum}\of{\mydata}
                \let\yval\pgfplotsretval
                \node[
                    at={(normalized axis cs:%
                            \pgfkeysvalueof{/data point/x},%
                            \yval)%
                    },
                    anchor=south,
                ]
                {
                    \pgfmathprintnumber{\yval}
                };
            },
        },
    }

\pgfplotsset{
    percentage ticks/.code={
        \pgfplotsset{
            nodes near coords={
                \pgfkeys{
                    /pgf/fpu=true,  % PGFplots uses floating points internally
                    /pgf/fpu/output format=fixed
                    %/pgf/number format/.cd,fixed,fixed zerofill, precision=2
                }
                \pgfmathparse{\pgfkeysvalueof{/data point/x}}
                \pgfmathtruncatemacro\xidx{\pgfmathresult}
                \pgfplotstablegetelem{\xidx}{sum}\of{\mydata}
                \pgfmathparse{100*\pgfplotspointmeta/\pgfplotsretval}
                \let\percent\pgfmathresult
                \ifthenelse{\lengthtest{\percent pt=100 pt}}{
                    \pgfmathprintnumber[fixed,fixed zerofill, precision=0]{\percent}\%
                }{
                    \ifthenelse{\lengthtest{\percent pt<10 pt}}{
                        \pgfmathprintnumber[fixed,fixed zerofill, precision=2]{\percent}\%
                    }{
                        \pgfmathprintnumber[fixed,fixed zerofill, precision=1]{\percent}\%
                    }
                }
            },
            every node near coord/.append style={yshift=1.2ex,xshift=-1.7ex,font=\footnotesize} %anchor=west,xshift=-7.2ex}
        }
    }
}

\begin{axis} [
    xtick=data,
    flexible xticklabels from table={data.txt}{gran}{col sep=comma},
    xticklabel style={text height=1.5ex},
    nodes near coords,
    percentage ticks
]
    \addplot table[x expr=\coordindex, y=two]{\mydata};
    \addplot [show sum on top] table[x expr=\coordindex, y=three]{\mydata}; % Note that "show sum on top" should be applied to the plot which always is different from 0
    \addplot table[x expr=\coordindex, y=four]{\mydata};
    \legend{Prefetch, Execute, Optimized}
\end{axis}

I'm afraid this is not a very minimal working example, but it shows the ideas:

Calculating the percentage by getting the x index, which is then used to find the sum in the table, and then taking the y-value of the node divided by the sum.

Adding the sum atop the bar is somewhat alike. For one item in each x-axis (I choose the "two" plot, cause that is the one which always has non-zero values), you simply create a node with the (x=x, y=sum value) coordinates. Apparantly zero-values are ignored and I am not sure how to deal with that. Worth noting as well is that I only modified the snippets, as they are originally from other threads here on tex SE.

Related Solutions

[Tex/LaTex] Stacked bar-plot: display total value

To only generate one label per bar stack, add the nodes near coords option to the last series only, using \addplot +[nodes near coords] coordinates....

In order to place the labels at the end of the bars, and not in the middle of the bar, starting with PGFPlots version 1.9 you have to add nodes near coords xbar stacked configuration/.style={} before calling xbar stacked.

To print the sum of the bar stacks, instead of the value of the last bar part, starting with PGFPlots version 1.9 you have to add point meta=x to your \addplot options.

Code for PGFPlots >= 1.9:

\documentclass[border=5mm]{standalone}
\usepackage{pgfplots}
\pgfplotsset{compat=1.12}

\pgfplotsset{testbar/.style={
    nodes near coords xbar stacked configuration/.style={},
    xbar stacked,
    width=.8\textwidth,
    axis y line*= none, axis x line*= bottom,
    xmajorgrids = true,
    xmin=0,xmax=70,
    ytick = data, yticklabels = {C1,C2},
    tick align = outside, xtick pos = left,
    bar width=6mm, y=8mm,
    enlarge y limits={abs=0.625},% 0.5 + 0.5*(y - bar width)/y [TeX.sx #47995]
}}

\begin{document}

\begin{tikzpicture}
    \begin{axis}[testbar] 
        \addplot coordinates{(2.68,1) (2.94,0)};
        \addplot coordinates{(26.63,1) (31.67,0)};
        \addplot +[
            point meta=x,
            nodes near coords,
            nodes near coords align={anchor=west},
            every node near coord/.append style={
                black,
                fill=white,
                fill opacity=0.75,
                text opacity=1,
                outer sep=\pgflinewidth % so the label fill doesn't overlap the plot
            }
        ]coordinates{(6.2,1) (9.91,0)};
    \end{axis}
\end{tikzpicture}

\end{document}

Code for PGFPlots < 1.9:

\documentclass[border=5mm]{standalone}
\usepackage{pgfplots}

\pgfplotsset{testbar/.style={
    xbar stacked,
    width=.8\textwidth,
    axis y line*= none, axis x line*= bottom,
    xmajorgrids = true,
    xmin=0,xmax=70,
    ytick = data, yticklabels = {C1,C2},
    tick align = outside, xtick pos = left,
    bar width=6mm, y=8mm,
    enlarge y limits={abs=0.625},% 0.5 + 0.5*(y - bar width)/y [TeX.sx #47995]
}}

\begin{document}

\begin{tikzpicture}
    \begin{axis}[testbar] 
        \addplot coordinates{(2.68,1) (2.94,0)};
        \addplot coordinates{(26.63,1) (31.67,0)};
        \addplot +[
            nodes near coords,
            nodes near coords align={horizontal},
            every node near coord/.append style={
                black,
                fill=white,
                fill opacity=0.75,
                text opacity=1,
                outer sep=\pgflinewidth % so the label fill doesn't overlap the plot
            }
        ]coordinates{(6.2,1) (9.91,0)};
    \end{axis}
\end{tikzpicture}

\end{document}

[Tex/LaTex] plotting two time series with bounds

This happens because PGFPlots only uses one "stack" per axis: You're stacking the second confidence interval on top of the first. The easiest way to fix this is probably to use the approach described in "Is there an easy way of using line thickness as error indicator in a plot?": After plotting the first confidence interval, stack the upper bound on top again, using stack dir=minus. That way, the stack will be reset to zero, and you can draw the second confidence interval in the same fashion as the first:

\documentclass{standalone}
\usepackage{pgfplots, tikz}

\usepackage{pgfplotstable}

\pgfplotstableread{
temps   y_h y_h__inf    y_h__sup    y_f y_f__inf    y_f__sup    

1   0.237340    0.135170    0.339511    0.237653    0.135482    0.339823    
2   0.561320    0.422007    0.700633    0.165871    0.026558    0.305184    
3   0.694760    0.534205    0.855314    0.074856    -0.085698   0.235411    
4   0.728306    0.560179    0.896432    0.003361    -0.164765   0.171487    
5   0.711710    0.544944    0.878477    -0.044582   -0.211349   0.122184    
6   0.671241    0.511191    0.831291    -0.073347   -0.233397   0.086703    
7   0.621177    0.471219    0.771135    -0.088418   -0.238376   0.061540    
8   0.569354    0.431826    0.706882    -0.094382   -0.231910   0.043146    
9   0.519973    0.396571    0.643376    -0.094619   -0.218022   0.028783    
10  0.475121    0.366990    0.583251    -0.091467   -0.199598   0.016664    
}{\table}

\begin{document}
\begin{tikzpicture}

    \begin{axis}
    % y_h confidence interval
    \addplot [stack plots=y, fill=none, draw=none, forget plot]   table [x=temps, y=y_h__inf]   {\table} \closedcycle;
    \addplot [stack plots=y, fill=gray!50, opacity=0.4, draw opacity=0, area legend]   table [x=temps, y expr=\thisrow{y_h__sup}-\thisrow{y_h__inf}]   {\table} \closedcycle;
    % subtract the upper bound so our stack is back at zero
    \addplot [stack plots=y, stack dir=minus, forget plot, draw=none] table [x=temps, y=y_h__sup] {\table};

    % y_f confidence interval
    \addplot [stack plots=y, fill=none, draw=none, forget plot]   table [x=temps, y=y_f__inf]   {\table} \closedcycle;
    \addplot [stack plots=y, fill=gray!50, opacity=0.4, draw opacity=0, area legend]   table [x=temps, y expr=\thisrow{y_f__sup}-\thisrow{y_f__inf}]   {\table} \closedcycle;

    % the line plots (y_h and y_f)    
    \addplot [stack plots=false, very thick,smooth,blue]  table [x=temps, y=y_h]   {\table};
    \addplot [stack plots=false, very thick,smooth,blue]  table [x=temps, y=y_f]   {\table};
    \end{axis}

\end{tikzpicture}
\end{document}

Best Answer

Related Solutions

[Tex/LaTex] Stacked bar-plot: display total value

[Tex/LaTex] plotting two time series with bounds

Related Question