LAUSR

The work presents the application of compositional data methodology to analytical results, taking as an example the study of the volatile profiles of green Spanish-style table olives according to cultivars and production areas. For this purpose, the volatile compounds (analysed by GC-MS and expressed as percentages of the total area) were considered as a compositional data set in the Simplex space and, as a result, analysed by their specific new statistical tools. Application of compositional exploratory tools (variation array, tertiary graphs, biplots, or coda-dendrogram) allowed differentiating cultivars and production areas based on their volatile profiles. Also, the application of Cluster and Principal Component analysis to the ilr transformed values (coordinates), following the new methodology, led to more realistic results than the formally incorrect implementation of the standard multivariate analysis (developed for data from the Euclidean space) to percentages (data in the Simplex). Therefore, the work presents a novel consideration of the volatile profiles of table olives as compositional data and shows their proper analysis by statistical tools specifically developed for them.