Data Sets are a group of parameter names and their values. These values can be quickly copied to one job, a group of jobs based on a filter or all jobs within a project. When you apply a Data Set then all you do is update the default value of a parameter. If the parameter does not exist in a job then it is not created. Both the run time values and the design time values are copied to a job. Most developers do not know that parameters are stored twice. Otherwise you would need to compile the jobs to update the run time values.
In order to update the default values of parameters in a project then you need to right click on the Data Set you want to use. There is a menu option to Select Data Set to Apply. Next you need to find the job or project you want to update then right click. There is a menu option which pops up called Apply Data Set. Select this menu option. It should tell you how many jobs and parameters were updated.
This is very fast and probably the most useful part of the product. Typical data sets are for development, test and production versions of a project. The data set has a field called PROJECT_TYPE the values for this field are DEV, TEST and PROD. This field must match the same field in a project definition. If not then you get an error saying the PROJECT_TYPE does not match. This will save you from updating the wrong values to the wrong project.
More Detail