Data Crunching is the preliminary step of automated, intelligent data processing, where large amounts of information are handled. The data is extracted, sorted and structured so that algorithms and programs can run robustly. Data crunching involves the retrieval of unstructured, raw data, the conversion of information from one format to another, e.g. text in XML records, and the correction of errors, e.g. typos. Usually, the tasks include reconciling datasets, finding inaccuracies, evaluating the overall quality of the data, deduplication, and column segmentation.

Depending on the context, data scientists use various tools and algorithms: in the past Excel and now, languages ​​such as Java, Python or R.