Fundamentals of Statistics contains material of various lectures and courses of H. Lohninger on statistics, data analysis and here for more.

Autocorrelation and Scatter Plots

Autocorrelation can easily be shown when one tries to plot the original value against a time-shifted copy of it. Look, for example, at the population of rabbits. The population shows a regular fluctuation within a period of 10 years.

Plotting the rabbit data against itself with no time shift, of course results in a correlation coefficient of 1.0. A time shift of one year already produces a considerably lower correlation coefficient of 0.63.

You should take a closer look at the relationship between the autocorrelation function and the scatter plots by starting the  DataLab .

Last Update: 2012-10-08