In this paper three topics concerning data clustering are discussed. One is an association found in supervised classification and clustering, whereby new techniques of clustering can be developed some of which are described here. Next two are concerned with risk mining. The consideration of associations between the above two classes of methods is related to the second topic in which clustering in the presence of rough sets is considered. The third is related to mining actual data on a risk issue, where the method of fuzzy clustering is applied. Throughout these considerations, we focus more on a research perspective, i.e., possibilities in near future rather than an established method or algorithm.