Over the past year I discovered LaTeX.

As in latex gloves?

No. It's pronounced "Lah-tech" or "Lay-tech". For the uninitiated LaTeX is a programming language used to produce documents. It can be used for books, articles, posters, presentation and many more things.

So a bit like Word then?

Nothing like word or any other word processing programming. Word processing programmes are great when you have nothing but text. However, I'm sure that everyone has experienced the annoyance of trying to put an image into a word document then finding it disappears onto another page or behind your text. Word and similar programmes are "What you see to what you get" (WYWISYG). LaTeX is "What you mean is what you get".

What do you mean?

If in Word I want to put an image 2.54cm from the right hand edge of the paper and 5.53cm from the top I may succeed to start with. However, once I add another image or some text there is no knowing whether it will stay there or not. In LaTeX it will stay exactly where I told it to.

Who should use LaTeX?

The great thing about LaTeX is that you can add packages to the basic installation. Packages can deal with mathematics, make graphs, posters, define colours, make books. If you use equations, graphs etc. you may find it worthwhile. also great for phonetics, ancient languages and languages using less commonly used alphabets.

Sounds a bit complicated...

Yes it is. I had a few false starts. There is quite a good introduction on wikibooks. At some point I plan to write a very basic introduction myself. The LaTeX project page is also a good place to start.

So everyone loves it then?

No, but I think its beautiful.

Show me an example.

See my online statistics book (preview).

How much does it cost?

Nothing

Amended 10 March 2016 (corrections/ update made)

I came across Anscombe’s Quartet on Wikipedia recently. I must confess to not having seen it before and don’t recall seeing it in any introductory statistics books.

The Anscombe’s Quartet is a conceptually and graphically clear way of showing the importance of graphs in statistical analysis. Each of the 11 pairs of observations have the same, x mean, y mean, x variance, y variance, correlation co-efficient and regression equation, though each have very different distributions. They clearly demonstrate the impact of outliers and how non-linear relationships can be identified.

Citation:

F. J. Anscombe (1973) Graphs in Statistical Analysis The American Statistician , Vol. 27, No. 1 (Feb., 1973), pp. 17-21

Article Stable URL: http://www.jstor.org/stable/2682899 (Not open access)

LaTeX code below.

\documentclass{article}
\usepackage{pgfplots}
\usepackage{pgfplotstable}
\pgfplotsset{compat=1.7}
\usepackage{amssymb, amsmath}
\usepackage{subcaption}
\begin{document}
\begin{figure}
\caption{Anscombe's quartet is a good demonstration why a scatterplot is so valuable, prior to calculating regression equations and correlation co-efficients. In all four cases the $x's$ have a mean of 9, and variance of 11. The mean of all the $y's$ is 7.5, and a variance 4.125. The correlation co-efficient of each is 0.816 and the linear regression line is $y=3+0.5x $}
\begin{subfigure}{.45 \textwidth}
\centering
\caption{Normal linear relationship}
\begin{tikzpicture}
\begin{axis} [width=5cm, height=5cm, xlabel=X1, ylabel=Y1]
\addplot[scatter, only marks, mark=x, mark size=4pt]
coordinates
{
(10, 8.04)
(8.0, 6.95)
(13, 7.58)
(9, 8.81)
(11, 8.33)
(14, 9.96)
(6, 7.24)
(4, 4.26)
(12, 10.84)
(7, 4.82)
(5, 5.68)
};
\addplot[scatter, mark=.]
coordinates
{
(0, 4.1)
(20, 12.5)
};
\end{axis}
\end{tikzpicture}
\end{subfigure}
\begin{subfigure}{.45 \textwidth}
\centering
\caption{Relationship clear, but not linear}
\begin{tikzpicture}
\begin{axis}[width=5cm, height=5cm, xlabel=X2, ylabel=Y2]
\addplot[scatter, only marks, mark=x, mark size=4pt]
coordinates
{
(10, 9.14)
(8.0, 8.14)
(13, 8.74)
(9, 8.77)
(11, 9.26)
(14, 8.10)
(6, 6.13)
(4, 3.1)
(12, 9.13)
(7, 7.26)
(5, 4.74)
};
\addplot[scatter, mark=.]
coordinates
{
(0, 4.1)
(20, 12.5)
};
\end{axis}
\end{tikzpicture}
\end{subfigure}
\
\begin{subfigure}{.45 \textwidth}
\centering
\caption{Clear linear relationship, but one outlier offsets the regression line}
\begin{tikzpicture}
\begin{axis} [width=5cm, height=5cm, xlabel=X3, ylabel=Y3]
\addplot[scatter, only marks, mark=x, mark size=4pt]
coordinates
{
(10, 7.46)
(8.0, 6.77)
(13, 12.74)
(9, 7.11)
(11, 7.81)
(14, 8.84)
(6, 6.08)
(4, 5.39)
(12, 8.15)
(7, 6.42)
(5, 5.73)
};
\addplot[scatter, mark=.]
coordinates
{
(0, 4.1)
(20, 12.5)
};
\end{axis}
\end{tikzpicture}
\end{subfigure}
\begin{subfigure}{.45 \textwidth}
\centering
\caption{Clear relationship, but one outlier puts the regression line at 45 degrees to the other 10 observations}
\begin{tikzpicture}
\begin{axis} [width=5cm, height=5cm, xlabel=X4, ylabel=Y4]
\addplot[scatter, only marks, mark=x, mark size=4pt]
coordinates
{
(8, 6.58)
(8.0, 5.76)
(8, 7.71)
(8, 8.84)
(8, 7.04)
(8, 5.26)
(19, 12.5)
(8, 5.56)
(8, 7.91)
(8, 6.89)
(8, 6.89)
};
\addplot[scatter, mark=.]
coordinates
{
(0, 4.1)
(20, 12.5)
};
\end{axis}
\end{tikzpicture}
\end{subfigure}
\end{figure}
\end{document}


Amended: 22 February 2016: There were a couple of errors in the code which I have now fixed. The previous code omitted the need for the xcolor package and the some coding items symbols (notably '\' were missing).
I have been searching the internet on how to produce a normal distribution curve in LaTeX with the standard deviations marked. I wasn't able to find exactly what I wanted, but got some good clues here.  This code uses the pgfplots package. The code should work 'as is'.


\documentclass{article}
\usepackage{pgfplots}
\usepackage{amssymb, amsmath}
\usepackage{tikz}
\usepackage{xcolor}
\pgfplotsset{compat=1.7}
\begin{document}
\pgfmathdeclarefunction{gauss}{2}{\pgfmathparse{1/(#2*sqrt(2*pi))*exp(-((x-#1)^2)/(2*#2^2))}%
}
\begin{tikzpicture}
\begin{axis}[no markers, domain=0:10, samples=100,
axis lines*=left, xlabel=Standard deviations, ylabel=Frequency,,
height=6cm, width=10cm,
xtick={-3, -2, -1, 0, 1, 2, 3}, ytick=\empty,
enlargelimits=false, clip=false, axis on top,
grid = major]
\addplot [fill=cyan!20, draw=none, domain=-3:3] {gauss(0,1)} \closedcycle;
\addplot [fill=orange!20, draw=none, domain=-3:-2] {gauss(0,1)} \closedcycle;
\addplot [fill=orange!20, draw=none, domain=2:3] {gauss(0,1)} \closedcycle;
\addplot [fill=blue!20, draw=none, domain=-2:-1] {gauss(0,1)} \closedcycle;
\addplot [fill=blue!20, draw=none, domain=1:2] {gauss(0,1)} \closedcycle;
\addplot[] coordinates {(-1,0.4) (1,0.4)};
\addplot[] coordinates {(-2,0.3) (2,0.3)};
\addplot[] coordinates {(-3,0.2) (3,0.2)};
\node[coordinate, pin={68.2\%}] at (axis cs: 0, 0.4){};
\node[coordinate, pin={95\%}] at (axis cs: 0, 0.3){};
\node[coordinate, pin={99.7\%}] at (axis cs: 0, 0.2){};
\node[coordinate, pin={34.1\%}] at (axis cs: -0.5, 0){};
\node[coordinate, pin={34.1\%}] at (axis cs: 0.5, 0){};
\node[coordinate, pin={13.6\%}] at (axis cs: 1.5, 0){};
\node[coordinate, pin={13.6\%}] at (axis cs: -1.5, 0){};
\node[coordinate, pin={2.1\%}] at (axis cs: 2.5, 0){};
\node[coordinate, pin={2.1\%}] at (axis cs: -2.5, 0){};
\end{axis}
\end{tikzpicture}
\end{document}