A proposal has been made to create an application which displays a map of Scotland and, when clicked, synthesizes speech approximating the dialect of the location clicked on, regardless of whether or not data has been collected for that precise area. This is to be accomplished by creating HMM voices for a number of locations across Scotland and interpolating appropriately between these voices at locations specified by the user. This report serves as a proof of concept for this project. It addresses two concerns: one, demonstrating that the Voices of Young Scots corpus, which collects recordings of youth across Scotland, is a suitable corpus for training these models; and two, building an extensible framework for the application and testing it using a set of voices with known geographical information. Both of these concerns are demonstrated to be surmountable, and some discussion is made of future work.