Latex: Two bar charts, the same legend

Question

I have two bar charts side by side, and they share the same legend, I have tried to use \addlegendentry{}, but I find my self with a repeated legend!

Here is the code, what should I do (I tried two many solution online but nothing work):

\begin{figure*}[h!]
\begin{adjustbox}{center}
\pgfplotsset{width=0.93\textwidth,height=0.4\textheight,compat=1.16}
\begin{tikzpicture}
\begin{axis}[
   legend style={at={(0.5,-0.09)},
      anchor=north,legend columns=-1},
    symbolic x coords={Precision,Recall,F1-Score,Accuracy},
    major tick length=0cm,
    xtick=data,
    ymin=75.0,
     enlarge x limits=0.2,
    enlarge y limits={upper,value=0.2},
    nodes near coords,
    ybar,
    every node near coord/.append style={rotate=90, anchor=west},
    bar width = 8pt,
]

\addplot
coordinates {(Precision,86.6) (Recall,86.0) (F1-Score,86.3) (Accuracy,98.4)};
\addlegendentry{KNN}

\addplot
coordinates {(Precision,87.6) (Recall,86.2) (F1-Score,86.5) (Accuracy,97.5)};
\addlegendentry{Neural Networks}

\addplot
coordinates {(Precision,85.9) (Recall,86.3) (F1-Score,86.1) (Accuracy,98.6)};
\addlegendentry{Decision Tree}

\addplot
coordinates {(Precision,87.1) (Recall,86.8) (F1-Score,87.0) (Accuracy,99.2)};
\addlegendentry{Random Forest}

\addplot
coordinates {(Precision,85.5) (Recall,82.8) (F1-Score,83.7) (Accuracy,97.4)};
\addlegendentry{SVM}

\addplot
coordinates {(Precision,77.9) (Recall,77.7) (F1-Score,75.9) (Accuracy,87.8)};
\addlegendentry{Naive Bayes}

\end{axis}

\end{tikzpicture}% NO EMPTY LINE HERE!!!!
\begin{tikzpicture}

\begin{axis}[
   legend style={at={(0.5,-0.09)},
      anchor=north,legend columns=-1},
    symbolic x coords={FPR,FNR},
    major tick length=0cm,
    xtick=data,
    ymin=0.0,
    enlarge x limits=0.5,
    enlarge y limits={upper,value=0.2},
    nodes near coords,
    ybar,
    every node near coord/.append style={rotate=90, anchor=west},
    bar width = 15pt,
]

\addlegendentry{KNN}
\addplot
coordinates {(FPR,7.90) (FNR,14.0)};

\addlegendentry{Neural Networks}
\addplot
coordinates {(FPR,3.45) (FNR,13.75)};

\addlegendentry{Decision Tree}
\addplot
coordinates {(FPR,0.39) (FNR,13.63)};

\addlegendentry{Random Forest}
\addplot
coordinates {(FPR,0.31) (FNR,13.13)};

\addlegendentry{SVM}
\addplot
coordinates {(FPR,5.38) (FNR,17.11)};

\addlegendentry{Naive Bayes}
\addplot
coordinates {(FPR,3.63) (FNR,22.19)};

\end{axis}

\end{tikzpicture}
\end{adjustbox}
\caption{Left - No Interaction. Right - Interaction} \label{fig:M}
\end{figure*}

You should have a look at this answer https://tex.stackexchange.com/a/326388/141947 — BambOo, May 31 '20 at 12:53
Thank u sir for your comment, but i didn't understand the answer in this question .. i'm newbie with Latex .. so i dont know what to do .. — Phd student, May 31 '20 at 13:00
Welcome to TeX.SX! I guess the simplest way to achieve that would be to use legend to name in the axis options. Have a look for that in the PGFPlots manual section 4.9.7 (v1.17). — Stefan Pinnow, May 31 '20 at 13:52

score 3 · Answer 1 · answered May 31 '20 at 14:01

Here is a proposal using simplified code based on the groupplots library

\documentclass[tikz]{standalone}
\usepackage{pgfplots}
\usepgfplotslibrary{groupplots}

\pgfplotsset{compat=1.16}


\begin{document}
\begin{tikzpicture}
    \begin{groupplot}[group style={group size=2 by 1},width=12cm,height=8cm]
        \nextgroupplot[
            %legend style={at={(mygroup.south)},anchor=north,legend columns=-1},
            symbolic x coords={Precision,Recall,F1-Score,Accuracy},
            major tick length=0cm,
            xtick=data,
            ymin=75.0,
            enlarge x limits=0.2,
            enlarge y limits={upper,value=0.2},
            nodes near coords,
            ybar,
            every node near coord/.append style={rotate=90, anchor=west},
            bar width = 8pt,
        ]

        \addplot coordinates {(Precision,86.6) (Recall,86.0) (F1-Score,86.3) (Accuracy,98.4)};
        \addplot coordinates {(Precision,87.6) (Recall,86.2) (F1-Score,86.5) (Accuracy,97.5)};
        \addplot coordinates {(Precision,85.9) (Recall,86.3) (F1-Score,86.1) (Accuracy,98.6)};
        \addplot coordinates {(Precision,87.1) (Recall,86.8) (F1-Score,87.0) (Accuracy,99.2)};
        \addplot coordinates {(Precision,85.5) (Recall,82.8) (F1-Score,83.7) (Accuracy,97.4)};
        \addplot coordinates {(Precision,77.9) (Recall,77.7) (F1-Score,75.9) (Accuracy,87.8)};
        %\legend{KNN,Neural Networks,Decision Tree,Random Forest,SVM,Naive Bayes}

        \nextgroupplot[
            legend style={at={(-\pgfkeysvalueof{/pgfplots/group/horizontal sep}/2,-0.1)},
                    anchor=north,legend columns=-1},
            symbolic x coords={FPR,FNR},
            major tick length=0cm,
            xtick=data,
            ymin=0.0,
            enlarge x limits=0.5,
            enlarge y limits={upper,value=0.2},
            nodes near coords,
            ybar,
            every node near coord/.append style={rotate=90, anchor=west},
            bar width = 15pt,
        ]

        \addplot coordinates {(FPR,7.90) (FNR,14.0)};
        \addplot coordinates {(FPR,3.45) (FNR,13.75)};
        \addplot coordinates {(FPR,0.39) (FNR,13.63)};
        \addplot coordinates {(FPR,0.31) (FNR,13.13)};
        \addplot coordinates {(FPR,5.38) (FNR,17.11)};
        \addplot coordinates {(FPR,3.63) (FNR,22.19)};


        \legend{KNN,Neural Networks,Decision Tree,Random Forest,SVM,Naive Bayes}
    \end{groupplot}
\end{tikzpicture}

\end{document}

score 1 · Answer 2 · answered May 31 '20 at 14:01

Thanks u so much sir for answering me. i found an easy solution

\begin{figure*}[h!]
\begin{adjustbox}{center}
\pgfplotsset{width=0.93\textwidth,height=0.4\textheight,compat=1.16}
\begin{tikzpicture}   
\begin{axis}[
    legend columns=-1,
    legend entries={KNN, Neural Networks, Decision Tree, Random  Forest, SVM, Naive Bayes},
    legend to name=named,
    legend style={at={(0.5,-0.09)},
    anchor=north,legend columns=-1},
    symbolic x coords={Precision,Recall,F1-Score,Accuracy},
    major tick length=0cm,
    xtick=data,
    ymin=75.0,
     enlarge x limits=0.2,
    enlarge y limits={upper,value=0.2},
    nodes near coords,
    ybar,
    every node near coord/.append style={rotate=90, anchor=west},
    bar width = 8pt,
]

\addplot 
coordinates {(Precision,86.6) (Recall,86.0) (F1-Score,86.3) (Accuracy,98.4)};

\addplot        
coordinates {(Precision,87.6) (Recall,86.2) (F1-Score,86.5) (Accuracy,97.5)};

\addplot        
coordinates {(Precision,85.9) (Recall,86.3) (F1-Score,86.1) (Accuracy,98.6)};

\addplot        
coordinates {(Precision,87.1) (Recall,86.8) (F1-Score,87.0) (Accuracy,99.2)};

\addplot        
coordinates {(Precision,85.5) (Recall,82.8) (F1-Score,83.7) (Accuracy,97.4)};

\addplot        
coordinates {(Precision,77.9) (Recall,77.7) (F1-Score,75.9) (Accuracy,87.8)};

\end{axis}
\end{tikzpicture}% NO EMPTY LINE HERE!!!! 
\begin{tikzpicture}  
\begin{axis}[
    legend style={at={(0.5,-0.09)},
    anchor=north,legend columns=-1},
    symbolic x coords={FPR,FNR},
    major tick length=0cm,
    xtick=data,
    ymin=0.0,
    enlarge x limits=0.5,
    enlarge y limits={upper,value=0.2},
    nodes near coords,
    ybar,
    every node near coord/.append style={rotate=90, anchor=west},
    bar width = 15pt,
]

\addplot 
coordinates {(FPR,7.90) (FNR,14.0)}; 

\addplot        
coordinates {(FPR,3.45) (FNR,13.75)};

\addplot 
coordinates {(FPR,0.39) (FNR,13.63)}; 

\addplot        
coordinates {(FPR,0.31) (FNR,13.13)};

\addplot 
coordinates {(FPR,5.38) (FNR,17.11)}; 

\addplot        
coordinates {(FPR,3.63) (FNR,22.19)};

\end{axis}
\end{tikzpicture}
\end{adjustbox}

\ref{named}
\caption{Left - No Interaction. Right - Interaction}
\label{fig:M}
\end{figure*}

It would be great if you could add an image of your result, so others have an easy way to see how your result looks like and can compare it to other answers. — Stefan Pinnow, May 31 '20 at 14:23

Latex: Two bar charts, the same legend

2 Answers2

Linked

Related