Exploratory Data Analysis Using Interactive Parallel Coordinates Plot Technique in an Epidemiological Study of Multidimensional Health Outcomes.

Aweke A. Mitku, University of KwaZulu-Natal

Background: We propose that parallel coordinates plots can be used to study multidimensional data particularly to explore the discovery of patterns across the variables. This can assist researchers from the health sciences to visualize their cohort data with interactive data analysis. Methods: In this paper, we demonstrate the exploration of multidimensional data with the use of brushing with color codes in parallel coordinates plots, assisting with the identification of relationships and patterns. The study used data from Mother and Child in the Environment birth cohort, for the period 2013 to 2017 retrospectively registered. Results: We have identified trends in the data that provide directions for further research, and illustrated thereby the potential of parallel coordinates plot to explore the patterns and relationships of annual prenatal oxides of nitrogen exposure with multidimensional birth outcomes. Conclusions: Brushing with color codes facilitated the detection of patterns of relationships in order to perform basic and advanced statistical model based analysis.