A Dataset is a collection of data extracted from Salesforce. 

For example...when you log in for the first time, Timmr will automatically load your Saleforce leads, contacts and accounts and then begin analyzing them for duplicates.  Once the analysis is completed, the Trimmr dashboard will display a Dataset for each object (along with the total number of records and duplicates in each).

Please keep in mind that the more records you have, the longer the initial analysis may take…but you will receive an email as soon it is finished.