The online-offline framework is a popular framework among others for clustering data stream. In this book, improving the online phase as well as the offline phase is one of our main considerations. For online phase, we retain the data summary information and information of the varying density clustering. Meanwhile, for offline phase, we recommend a novel algorithm of varying density clustering. The algorithm utilizes the online synopsis information to form the final clusters making an allowance for the data density distribution. Using different metrics of assessment, a series of exhaustive experimentations were conducted on the synthetic and real datasets. The outcomes substantiated that the proposed method is able to cluster data stream in varying density environments compared to the prevailing techniques. We conducted a thorough assessment on a dissimilar data set with different numbers of size, cluster and density. Our proposed method was established to have a better clustering quality, efficiency and scalability compared to the existing approaches.