A data transformation is a mathematical operation that is applied to a series of data points. It typically results in a new set of data points with a different shape. The transformed values are connected to the original data points by a formula. If the original data points were positive, for example, the logarithm transformation is applied. In addition, the square root transformation and the multiplicative inverse transformation are common.
Data transformation improves the interpretability of data. For example, fuel economy data, which usually appears as miles per gallon, can be transformed to liters per mile or kilometers per liter. This can result in a more easily-understood graph. However, not all data is easily interpreted, and sometimes the transformations are not appropriate for the analysis.
A good example of this is the Robust Multi-Array Average (RMA) normalization, a statistical procedure that has become popular for normalizing microarray data. The OBCS model also shows other types of data processing methods. The OBCS ontology includes terms related to data transformation and data analysis.
Data transformation is a statistical process that improves the quality of data and makes it easier to use by humans and computers. Using data that has been transformed can help to eliminate issues related to data consistency and performance.