[Tex/LaTex] Create boxplots from file

boxplotpgfplotstikz-pgf

My question is somewhat related to this thread. I would like
to create box plots from a file. The file (in my case a CSV) looks like this:

Upper,Lower,Min,Max,Median,Name
3,1,0,4,2,First
4,2,1,5,3,Second
...

There are multiple rows, each of which contains fields describing the data points for the box plot (median, position of the whiskers and the box boundaries) and the name of the data set. I would like to create a number of box plots from this data. The name should be added as a legend entry.

I have tried using a CSV reader:

\begin{tikzpicture}
  \begin{axis}[
    \csvreader[
      head to column names,
      separator=comma]
              {data.csv}{}% use head of csv as column names
              {
                \addplot+ [
                  boxplot prepared={
                    lower whisker=\Min,
                    lower quartile=\Lower,
                    median=\Median,
                    upper quartile=\Upper,
                    upper whisker=\Max,
                  },
                ] coordinates {};
              }
  \end{axis}
\end{tikzpicture}

This produces an empty plot though. My guess is that the macros (\Min and so on) do not work properly in the tikzpicture / axis environment?! (It works when I just dump out the values outside of the picture).

I also took a look at the thread mentioned above. The problem in this case is that only the first row of the table is plotted whereas I would like to plot all of them. This shortcoming is mentioned in the comments, but no extension of the approach is mentioned.

What is the best way to add the plots?

Best Answer

You can extend Jake's answer to Read boxplot prepared values from a table fairly easily, using \pgfplotsinvokeforeach.

The code below is based on Jake's answer, with just a few modifications:

a way of getting the number of rows from a table:
```
\pgfplotstablegetrowsof{\datatable}
\pgfmathtruncatemacro\TotalRows{\pgfplotsretval-1}
```
Subtract one, because row/column numbers start counting from 0 in indexing.
a loop instead of two different \addplots:
```
\pgfplotsinvokeforeach{0,...,\TotalRows}{ \addplot .. }
```
Note the addition of row=#1 in the boxplot options, in \pgfplotsinvokeforeach the loop variable is represented by #1.
added area legend to the \addplot options, otherwise the legend image is a complete (large) box plot.

added legend entry from table:

\pgfplotstablegetelem{#1}{name}\of\datatable
\addlegendentryexpanded{\pgfplotsretval}

\documentclass[crop=false]{standalone}
\usepackage{pgfplotstable}
\pgfplotsset{compat=1.8}
\usepgfplotslibrary{statistics}
\makeatletter
\pgfplotsset{
    boxplot prepared from table/.code={
        \def\tikz@plot@handler{\pgfplotsplothandlerboxplotprepared}%
        \pgfplotsset{
            /pgfplots/boxplot prepared from table/.cd,
            #1,
        }
    },
    /pgfplots/boxplot prepared from table/.cd,
        table/.code={\pgfplotstablecopy{#1}\to\boxplot@datatable},
        row/.initial=0,
        make style readable from table/.style={
            #1/.code={
                \pgfplotstablegetelem{\pgfkeysvalueof{/pgfplots/boxplot prepared from table/row}}{##1}\of\boxplot@datatable
                \pgfplotsset{boxplot/#1/.expand once={\pgfplotsretval}}
            }
        },
        make style readable from table=lower whisker,
        make style readable from table=upper whisker,
        make style readable from table=lower quartile,
        make style readable from table=upper quartile,
        make style readable from table=median,
        make style readable from table=lower notch,
        make style readable from table=upper notch
}
\makeatother


\pgfplotstableread{
    lw lq med  uq uw name
     5  7 8.5 9.5 10 first
     4  5 6.5 8.5 9.5 second
}\datatable


\begin{document}
\begin{tikzpicture}
\begin{axis}[boxplot/draw direction=y]
\pgfplotstablegetrowsof{\datatable}
\pgfmathtruncatemacro\TotalRows{\pgfplotsretval-1}
\pgfplotsinvokeforeach{0,...,\TotalRows}
{
  \addplot+[
  boxplot prepared from table={
    table=\datatable,
    row=#1,
    lower whisker=lw,
    upper whisker=uw,
    lower quartile=lq,
    upper quartile=uq,
    median=med
  },
  boxplot prepared,
  % to get a more useful legend
  area legend
  ]
  coordinates {};

  % add legend entry 
  \pgfplotstablegetelem{#1}{name}\of\datatable
  \addlegendentryexpanded{\pgfplotsretval}
}
\end{axis}
\end{tikzpicture}
\end{document}

Related Solutions

[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}

[Tex/LaTex] pgfplots – Create multiple plots from one table (rows 0–3, 4–7,…)

Yes, this is possible. The main problem, why your \foreach solution isn't working is, because the \n+3 isn't evaluated. You can test this by, e.g. replacing select coords between index={12}{15} with select coords between index={12}{12+3} in your "manual" solution, which will raise an error.

I think I also have addressed all your "additional (marginal) questions" except for the last two. Regarding 5) I don't like grid lines, but you already have found the solution so I think you can apply that yourself if you really want/need them. And Regarding 6) I am not sure if you really just meant to output a PDF of "the whole document" or maybe just create a PDF of the plot (alone).

For more details please have a look at the comments in the code.

% used PGFPlots v1.14
\documentclass[border=5pt]{standalone}
\usepackage{xstring}        % <-- needed for string manipulation
\usepackage{pgfplots}
\usepackage{pgfplotstable}
    \pgfplotsset{
        % use at least this `compat' level to improve placing of axis labels
        compat=1.3,
        % (almost the) opposite of `skip coords between index'
        % only use x values when the `\coordindex' lays between the two
        % arguments (including), else "skip" them by setting the value to "NaN"
        select coords between index/.style 2 args={
            x filter/.code={
                \pgfmathifthenelse{or(\coordindex<#1,\coordindex>#2)}{NaN}{x}
                \pgfmathresult
            },
        },
        table/col sep=comma,
        table/ignore chars=\%,
    }
    % define a command to create the legend entries
    % (give the row number to the argument)
    \newcommand{\CustomLegend}[1]{
        % store the "Benchmark" entry of the given row number in `\Benchmark'
        \pgfplotstablegetelem{#1}{"Benchmark"}\of{\table}
            \edef\Benchmark{\pgfplotsretval}
        % store the "Capacity" entry of the given row number in `\Capacity'
        \pgfplotstablegetelem{#1}{"Param: capacity"}\of{\table}
            \edef\Capacity{\pgfplotsretval}
        % create the legend entry
        \addlegendentryexpanded{
            % only use the part of `\Benchmark' between the given two strings
            \noexpand\StrBetween{\Benchmark}{impl.}{s.test}
            % append the capacity entry in brackets if the "Benchmark" entry
            % ends with the given string (else return an empty string)
            \noexpand\IfEndWith{\Benchmark}{ArrayList"}{%
                (\Capacity)
            }{}
        }
    }
\begin{document}
    \begin{tikzpicture}
            % read the table from the file
            \pgfplotstableread{lists-results.csv}{\table}
        \begin{axis}[
            title=Lists collections performance,
            xlabel=Objects,
            % changed y axis label and corresponding style
            ylabel=Add operations in million ops/s,
            % use an index for the data points and get the labels from the table
            xtick={0,...,10},
            xticklabels from table={\table}{"Param: objects"},
            y tick label style={
                ytick scale label code/.code={},
                scaled y ticks=base 10:-6,
            },
            % changed positioning of the legend
            legend pos=outer north east,
            % use left alignment for labels in legend
            legend cell align=left,
        ]
            \foreach \n in {0,4,...,12}{
                \addplot+[mark=o,select coords between index={\n}{\n+3}] table [
                % to avoid a numerical issue I have added "+0.1" to x
                % (otherwise for \n=16 the modulus would return 4, because
                % 16 would be something like 15.999)
                    x expr={int(mod(\coordindex + 0.1,4))},
                    y expr=\thisrow{"Score"}*\thisrow{"Param: objects"},
                ] {\table};
                    \CustomLegend{\n}
            }
        \end{axis}
    \end{tikzpicture}
\end{document}

Best Answer

Related Solutions

[Tex/LaTex] plotting two time series with bounds

[Tex/LaTex] pgfplots – Create multiple plots from one table (rows 0–3, 4–7,…)

Related Question