Methods used in workflow

A workflow is used to clarify data and arrange it so that it can better sense and can be used for analysis. Methods usually adopted are a combination of many choices. First used is simple parsing where misspelled data or syntax checking is done. A specific set of specifications are followed. Duplicated data is eliminated as well from the data set by the algorithm which sorts the data. This is followed by data transformation where the misfits are normalized to a set of specified minimum and maximum value. The data is mapped and then changed to fit in if it falls out by an algorithm. The algorithm also may use statistical method by analyzing the effect of data on mean, standard deviation, range and clustering (an algorithm). Unexpected values and erroneous data can be detected and placed for removal during workflow.

Statistical error values can be replaced by more normal values by augmentation algorithms which are quite complex. The results of the workflow process only pass after inspection of the data changes that are needed. They have to be inspected. If it is wrong they are manually changed back. After this phase if the data has been changed it may have to go to cycle 2 of cleansing and auditing, and ultimately the data will proceed to workflow stage 2.