Multivariate statistics applied to spinning and weaving instruments from Roman Venetia (Italy): some preliminary results

Statistical analysis is a useful research tool for the analysis of instruments related to textile activities, especially spindle whorls and loom weights. Scholars almost exclusively use the most basic statistical techniques, namely the univariate or bivariate methods, while the use of the multivariate analysis is rare. Yet, multivariate techniques are well suited for exploring the numerous variables of textile tools, as these methods take into account two or more variables simultaneously, allowing for synthesising, distinguishing and classifying the various components of such a complex dataset. The aim of this short contribution is to apply some of the most common multivariate techniques to specific samples of instruments used for spinning and weaving. These data have been collected within the scope of the project Pondera, carried out by the Department of Cultural Heritage of the University of Padua. After a brief introduction, the methods and outcomes of Principal Component Analysis, Correspondence Analysis and Cluster Analysis are discussed. The preliminary results partially confirm what previous works on the same data have already highlighted, but also provide new ideas concerning the correlations among textile tools from Roman Venetia which had not been fully explored thus far.