PhenomenaAssociater: Linking multi-domain spatio-temporal datasets

Abstract

This paper focuses on the demonstration of an analytics dashboard application for analyzing interesting spatio-temporal associations between anomalies across multiple spatio-temporal datasets, potentially from disparate domains, to find interesting hidden relationships. The proposed system is intended to analyze spatiotemporal data across multiple phenomena from disparate domains (for example traffic and weather) to identify interesting phenomena relationships by linking anomalies from each of these domain datasets. This web-based dashboard application developed in R Shiny [1] provides interactive visualizations to quantify the multi-domain associations. The application uses a novel framework of algorithms and quantification metrics to associate these anomalies across multiple domains using spatial and temporal proximity and influence metrics.