Discovering exactly what is in your data and how it might be useful for different analytic explorations is key to quickly identifying the value or potential use of a dataset.

This exploration process allows you to gain an understanding for the unique elements of the data such as value distributions and outliers to inform the transformation and analysis process.

Cleaning involves taking out data that might distort the analysis. A null value, for example, might bring an analytic package to a screeching halt; it may need to be replaced with a zero or an empty string.

Particular fields may need to be standardized by replacing the many different ways that a state for example might be written out -- such as CA, Cal and Calif -- with a single standard format.

Enriching allows you to augment the scope of your analysis by incorporating disparate internal or 3rd-party data into your analysis. This step includes executing common preparation tasks such as joins, unions or complex derivations.

Purchase transaction data, for example, might benefit from data associated with each customer's profile or historical purchase patterns.

Validating is the activity that surfaces data quality and consistency issues, or verifies that they have been properly addressed by applied transformations. Validations should be conducted along multiple dimensions.

At a minimum, assessing whether the values of an attribute/field adhere to syntactic constraints as well as distributional constraints.