One of the innovations data mining has brought to the field of data analysis is the discovery of patterns, such as association rules and subgroups. In the first part of the talk I will introduce and illustrate this concept of patterns.
Patterns also play another role in data analysis. Many ``modern'' types of data are complex in the sense that the data items themselves exhibit structure or patterns. Examples are large text corpera, multi-media databases, databases with millions of time-series and DNA/protein data. In the analysis of such data sets the ``within data item'' patterns can be very usefull. In the second part of the talk I will illustrate this using wavelets to describe the ``within'' patterns.