Toward an Anayist-Oriented Polystore Framework for Processing JSON Geo-Data

Geo-data sets represented as JSON documents are provided by many sources over the Internet. Consequently, many storage technologies are adopted for managing them, namely NoSQL databases such are MongoDB and ElasticSearch. To allow analysts to effectively perform complex analysis processes, such as integration and transformations possibly based on spatial properties, a framework that provides a unified view of different database systems based on different technology is necessary, so that analysts are unaware of specific characteristics of different database systems. In other words, we follow a polystore approach. The presented framework, named J-CO, provides a query language and an execution engine that, together, enable analysts to specify complex transformation processes on JSON geo-data sets. High-level operator natively support spatial operations, and completely hide technical details of the integrated database systems. This way, a novel data store component, named J-CO-DS, can be seamlessly introduced, to overcome limitations characterizing other NoSQL database systems. The paper presents the framework and the novel features (w.r.t. our previous works); then, it shows the effectiveness of the framework itself through an example.