Big Data for the Social Good: The Drought Early-Warning Experience Report

Abstract

Data deluge is growing exponentially, but the consumption of the data is not growing at the same pace. DataOps is an emerging family of techniques and tools that harnesses the potential of data continuously whilst incrementally using complex cloud systems orchestration techniques. This paper offers a proof-of-concept implementation of a DataOps pipeline for the social good. Specifically, we prototype—using a combined field-lab Action Research and Design Science approach—a DataOps system to incrementally and iteratively mitigate the devastating effects of droughts for high-risk areas. The context of our study is a game reserve in the Waterberg area in the province Limpopo in South Africa. The objective of this paper is threefold: to (1) develop and study a proof of concept for DataOps, by (2) exploring the applicability of individual software components in a complex and large-scale continuous pipeline, and, finally (3) elaborate on the spatial classification of such components in a new-frontier Drought Early-Warning System (DEWS). As a result, we offer an overview of the challenges and opportunities laid bare by our experimentation in a complex societal scenario while combining Artificial Intelligence and DataOps technologies. We conclude that a combined model of local, regional, and global data performs best on all tests within a stakeholder-acceptable timeframe.