A Classification of Pure Malt Scotch Whiskies

Abstract

Single-malt Scotch whiskies are produced by 109 distilleries in Scotland. The layman may wonder what are the major types of single malts that can be recognized, and what are their chief characteristics and best representatives, whether there is a geographical component in that classification and whether the various categories of characteristics lead to the same classification. This paper provides an answer to these questions, applying an array of statistical methods to a database derived from a connoisseur's description of these liquors. The taster's literary descriptions of Scotches were turned into a numerical database (109 Scotches x68 binary variables). A first classification was produced by distance computation and hierarchical clustering. Since it was significantly related to the regions of Scotland, a second classification was computed with a spatial contiguity constraint, to divide Scotland into regions where Scotches are homogeneous in their organoleptic characteristics. To explore the congruence of the categories of characteristics, the five databases corresponding to nose, colour, body, palate and finish characteristics were compared by using statistical tests of significance: among the raw data tables, among distance matrices and among classifications derived from these distance matrices. Most types of characteristic lead to congruent results, despite the loss of information that occurs