How to add a jitter to a scatter plot with pgfplots

Question

I have the following scatter plot:

\begin{tikzpicture}
  \begin{axis}[
    enlarge x limits=0.02,
    xmin=0,
  ]
    \addplot+[only marks,mark size=1pt] table[x=rmsd,y=seq_score_min] {\featuretable};
  \end{axis}
\end{tikzpicture}

I would like to add a jitter to the marks, i.e. some noise so that one can get a better feeling about the number of marks can would otherwise be placed above each other. In this particular case this is needed for the rightmost values.

Answer 1

You can use an x filter to add a bit of random noise to each x coordinate. I've defined a new jitter style that takes an optional argument to control the amount of noise:

\documentclass{standalone}
\usepackage{pgfplots,pgfplotstable}
\pgfplotstableread{
X   Y
1   2
1   2.1
1   2.05
1   2
1   2.2
1   2.15
1   2.25
1   2.1
1   2.025
1   2.125
1   2.0525
1   2.25
1   2.225
1   2.1525
1   2.2525
1   2.125
}\datatable

\pgfplotsset{
    jitter/.style={
        x filter/.code={\pgfmathparse{\pgfmathresult+rnd*#1}}
    },
    jitter/.default=0.1
}

\begin{document}
\begin{tikzpicture}
\begin{axis}[only marks,ymin=0,ymax=4,xmin=0,xmax=5]
\addplot +[jitter=0.3] table {\datatable};
\end{axis}
\end{tikzpicture}
\end{document}

Answer 2

To add jitter or noise I would rather do it in another program than TeX, however it can be done.

See this code which lets you add noise in the x or y direction.

\documentclass{article}
\usepackage{pgfplots}
\begin{document}
\begin{tikzpicture}
  \begin{axis}[domain=0:10,samples=100,
    my x filter/.style={%
        x filter/.code={%
            \def\tmpx{##1}%
            \pgfmathparse{\tmpx > 8 ? \tmpx+rnd*#1 : \tmpx}%
        }},
    my y filter/.style={%
        y filter/.code={%
            \def\tmpy{##1}
            \pgfmathparse{\tmpy > .75 ? \tmpy+rnd*#1 : \tmpy}
        }}]
    \addplot[blue,my x filter=.5] function {sin(x)};
    \addplot[red,my y filter=.1] function {cos(x)};
  \end{axis}
\end{tikzpicture}
\end{document}

This will produce (not exactly, due to the rand).