AStar : a modeling language for document-oriented geospatial data warehouses

Abstract

A Geospatial Data Warehouse (GDW) is an extension of a traditional Data Warehouse that includes geospatial data in the decision-making processes. Several studies have proposed the use of document-oriented databases in a GDW as an alternative to relational databases. This is due to the ability of non-relational databases to scale horizontally, allowing for the storage and processing of large volumes of data. In this context, modeling the manner in which facts and dimensions are structured is important in order to understand, maintain, and evolve the Document-oriented GDW (DGDW) through visual analysis. However, to the best of our knowledge, there are no modeling languages that support the design of facts and dimensions as referenced or embedded documents, partitioned into one or more collections. To overcome this lack, we propose Aggregate Star (AStar), a Domain-Specific Modeling Language for designing DGDW logical schemas. AStar is defined from a concrete syntax (graphical notation), an abstract syntax (metamodel), and static semantics (well-formedness rules). In order to describe the semantics of the concepts defined in AStar, translational semantics map the graphical notation to the metamodel and the respective code, to define the schema in MongoDB (using JSON Schema). We evaluate the graphical notation using Physics of Notations (PoN), which provides a set of principles for designing cognitively effective visual notations. This evaluation revealed that AStar is in accordance with eight of the nine PoN Principles, an adequate level of cognitive effectiveness. As a proof of concept, the metamodel and well-formedness rules were implemented in a prototype of Computer-Assisted Software Engineering tool, called AStarCASE. In its current version, AStarCASE can be used to design DGDW logical schemas and to generate their corresponding code in the form of JSON Schemas. Furthermore, we present a guideline that shows how to design schemas that have facts, conventional dimensions, and geospatial dimensions related as referenced or embedded documents, and partitioned into one or more collections. The guidelines also present good practices to achieve low data volume and low query runtime in a DGDW.