LET SAS CLEANSE YOUR DIRTY DATA

In an ideal world, every data set is complete, clean, and properly formatted. However, in real world situations, the data available to us is very rarely presented in this form. They may contain any number of problematic events such as outliers, duplicate observations, missing values, invalid character and numeric data values, as well as many other issues. Given the necessity that the data being examined is as complete and clean as possible, it is very important that these issues are addressed prior to any analysis. In this paper we describe various cases of dirty data and techniques to clean them. These techniques are explored within the context of SAS® 9.4 and presented in a way that would benefit beginning and moderate level SAS users.