Visual tools for the exploration of growth data in a cohort of kangaroo infants during their first year of life

摘要

The Fundación Canguro has been providing care for newborn premature infants in Colombia since 1994, helping them overcome the many issues related to their birth conditions. As part of their mission they have collected a longitudinal database of more than 33,000 children, recording their evolution during 23 years, with 47 measured variables. Having such rich data is both a blessing and a curse, as deriving insights from it is a daunting task. In this paper we present the results of a visual analytics project that has been enabling the analysts of the Fundación Canguro to derive insights in ways they couldn't do before. To support the project, we present a process model o framework that can be applied to any time based process where the results of an event are evaluated. For the problem at hand, the model considers the environment (before the event), pregnancy (development of the event), birth (the event), and growth (after the event). Using this model, a visual analytics tool was developed and implemented, and has been used by the analysts for the last months for deriving insights. The tool was developed applying Tamara Munzner's Visual Analytics framework [14] using advanced interactions and information visualizations, including a customized time series diagram. To demonstrate the value of the tool, we present two example case studies were analysts were able to conclude that educated women are less prone to have unwanted pregnancies, and that the implementation of a national law to increase maternity leave generated better feeding patterns on premature children in the Country.