STARE-based Interactive Analytics for Earth Science Big Data

Advanced observation and computation techniques allow Earth scientists to gain Earth Science Data characterized with enormous volume and diversity. Although existing work can effectively tackle vast volume of homogeneous data, it remains a grand challenge to achieve scalable analytics capabilities for heterogeneous data. We present an interactive analytics platform in support of Earth Science Big Data. We leverage a unifying indexing scheme, STARE, to partition and distribute datasets that are generated from different sources and with different data models. STARE has enabled us to place voluminous and diverse Earth Science data on different data management or store systems, such as SciDB and netCDF files. We demonstrate the scalable performance of our platform, as well as its enabled effective interactive data visualization and analysis capabilities.