Modelling and Designing Spatial and Temporal Big Data for Analytics

Abstract

The main purpose of this paper is to introduce a new approach with a new data model and architecture that supports spatial and temporal data analytics for meteorological big data applications. The architecture is designed with the recent advances in the field of spatial data warehousing (SDW) and spatial and temporal big data analytics. Measured meteorological data is stored in a big database (NoSQL database) and analyzed using Hadoop big data environment. SDW provides a structured approach for manipulating, analyzing and visualizing the huge volume of data. Therefore, the main focus of our study is to design a Spatial OLAP-based system to visualize the results of big data analytics for daily measured meteorological data by using the characteristic features of Spatial Online Analytical Processing (SOLAP), SDW, and the big data environment (Apache Hadoop). In this study we use daily collected real meteorological data from various stations distributed over the regions. Thus, we enable to do spatial and temporal data analytics by employing spatial data-mining tasks including spatial classification and prediction, spatial association rule mining, and spatial cluster analysis. Furthermore, a fuzzy logic extension for data analytics is injected to the big data environment.