Data mining refers to the search for useful statistical information in a large volume of data.

In multidimensional analysis reporting, the user knows what they are looking for: e.g. the breakdown of turnover by region. They don‟t yet know what the breakdown is, but they do know what it consists of, and that this knowledge is relevant to their work.

In data mining, the user is looking for hidden statistical information which they haven‟t yet been identified: trends, correlation, similarities, etc. Analysis of supermarket till receipts will typically show correlations between product purchases. While correlations some will have a simple explanations (razors & shaving gel), other will be more unexpected.

Such results need to be verified as the correlation observed may not be related to cause and effect but rather be just a coincidence, or the result of a hidden cause, both with phenomena‟s such as conjoint consequences.

Data mining tools semi-automatically search through large volumes of data for these correlations & statistical rules, sometimes using a visual reproduction of these rules, which will reveal, for example, clusters of points on an axial representation of data.