Document-Oriented Geospatial Data Warehouse: An Experimental Evaluation of SOLAP Queries

Geospatial Data Warehouse (GDW) is a repository of historical and geospatial data used in the decision-making process. This kind of system manages large volumes of data and supports Spatial On-Line Analytical Processing (SOLAP) queries. The use of NoSQL databases in the construction of GDW is still an unexplored topic, even though NoSQL databases present good performance and high scalability with low-cost hardware. In this context, we seek to identify the level of redundancy of geospatial data in Document-Oriented GDW (DGDW) that reduces the storage cost and increases the performance of SOLAP queries. Using the MongoDB database, we exhaustively define and investigate nine DGDW schemas, which have different levels of geospatial data redundancy in their dimensions. We performed an experimental evaluation of these schemas in a cluster structure, to analyze the data volume and the runtime of seven queries which simulate SOLAP operations in geospatial fields with different levels of redundancy and selectivity. Our experimental results indicate that the normalization of low-selectivity geospatial fields and the denormalization of high-selectivity geospatial fields are good strategies to reduce the storage cost and improve the performance of SOLAP queries. The results of our experimental evaluation are an important contribution because they can be used as a guide for construction of DGDW.