A high-speed gas chromatography method has been developed to separate the fatty acids in different milk fat fractions obtained via the dry fractionation deodorization process (TIRTIAUX method). Exploratory data techniques, namely cluster analysis and principal component analysis (PCA), were applied to gas chromatographic data for fatty acids composition in these milk fractions. Cluster analysis allowed the authors to visualize natural groupings corresponding to milk fat cracking achieved via the TIRTIAUX method. Principal component plots showed that six fractions are grouped in different classes, each group being clearly distinguished. Furthermore, a relationship was established between these classes and fatty acids through PCA analysis. A discriminant linear model for predicting milk fat classes from fatty acid composition was computed; 99 % of the criterion variable (milk fat class) variance was explained by the constructed model. Enrichment in unsaturated cis-fatty acids in the olein fractions and in the saturated and unsaturated trans-fatty acids in the stearin fractions was visualized from chemometric analysis. This differentiation of the milk fat fractions on the basis of fatty acids composition will enable food industries to use these fractions in both an efficient and safe way to enrich dairy products, or other foods, for healthy consumption.