[en] In this work, computational methods of finding clusters of multivariate data points were explored using principal component analysis (PCA), discriminate analysis (DA) and rough set analysis (RSA) methods. The variables were the concentrations of four natural isotopes and the texture characteristics of 100 sand samples from the coast of North Sinai, Egypt. Beach and dune sands are the two types of samples included. These methods were used to reduce the dimensionality of multivariate data and as classification and clustering methods. The results showed that the classification of sands in the environment of North Sinai is dependent upon the radioactivity contents of the naturally occurring radioactive materials and not upon the characteristics of the sand. The application of DA enables the creation of a classification rule for sand type and it revealed that samples with high negatively values of the first score have the highest contamination of black sand. PCA revealed that radioactivity concentrations alone can be considered to predict the classification of other samples. The results of RSA showed that only one of the concentrations of ²³⁸U, ²²⁶Ra and ²³²Th with ⁴⁰K content, can characterize the clusters together with characteristics of the sand. Both PCA and RSA result in the following conclusion: ²³⁸U, ²²⁶Ra and ²³²Th behave similarly. RSA revealed that one/two of them may not be considered without affecting the body of knowledge