As a part of this project, we are exploring a range of techniques applicable to multimedia as well as simulation data from scientific applications. For multimedia data, we are looking at techniques based on approximate pattern matching for image and video compression. We have derived theoretical results for performance of these schemes and developed a package -- 2D-PMC for image and video compression. Our work on analysis of simulation data centers around large scale particle dynamics systems evolved using approximate hierarchical methods. Compression techniques for such data use bounds on distortion radii to affine particles to quantized locations yielding high compression rates while maintaining error bounds. We are also exploring linear algebraic techniques based on matrix decompositions for analyzing data from both scientific and business applications.