Real world data is complex and multi-related among itself. Therefore, homogenous clustering will not be sufficient to analyze it. It will result lack of information. Such data should be dealt with k-partite clustering methods, because this method will include different types of nodes. At the end, it will give heterogeneous clusters which are effective to represent inter-partition relations as well as intra-partition ones. To exemplify, from a complex big relations of user- sentiment-issue, clusters may be extracted such that they contains users who uses similar sentiments to address same issues. I present a new algorithm, called STriCluster, which evaluates heterogeneous input data which contains rela- tions of three different types. Each relation called as hyperedge has three nodes from distinct types. It uses heuristics to mine clusters from these hyperedges. It follows a greedy approach to increase cover. Overlap of hyperedges among clusters are not allowed while a node can be part of many clusters. As a new feature, hyperedges has either positive or negative property. Our algorithm effectively handles negative property and sparseness of hyperedges while mining tri-partite clusters of hyperedges with positive properties. I will show its productivity via several experiments and results. Additionally, I will present its effectiveness via metrics.