Data cube is the core operator in data warehousing and OLAP. Its efficient computation, maintenance, and utilization for query answering and advanced analysis have been the subjects of numerous studies. However, there are some inherent problems that current techniques cannot handle well. Many kinds of critical semantic relationships among aggregate cells in a data cube are not captured. The huge size of the data cube limits its applicability as a means for semantic exploration. Although current compression approaches should help cut down the cube size, none of them retains the semantics of data cubes.

We developed quotient cube, a systematic approach to achieve efficacious data cube construction and exploration by semantic summarization and compression. The key idea behind a quotient cube is to create a summary structure by carefully partitioning the set of cells of a cube into classes such that cells in a class all have the same aggregate measure value and in addition, satisfy some desirable properties. Each class is a concise summary for all its cells. In addition, a quotient cube captures the roll-up/drill-down semantics between cells, present in the original cube, between its classes. QC-tree is a compact and efficient data structure to store and explore a quotient cube. We also developed efficient algorithms for answering queries, incremental maintenance against updates using QC-tree.

Based on our research achievements on quotient cube, we implemented SOQCET, a prototype data warehousing and OLAP system. SOQCET demonstrates the techniques of building a quotient cube, maintaining a quotient cube against update, using quotient cube to answer various queries and support cube navigation and exploration.