Radial and Network Visualizations
(Newborn Health)
— GCE Brazil (2020)
Background
Funded by the Bill & Melinda Gates Foundation, this project investigates the relationship between the conditional cash transfer program Bolsa Família (BF) designed to improve the health and education of children of a given generation in Brazil and health achievements of newborns in the next generation. The research team was composed by academics from the fields of health, economics and data science from 3 different universities: UFPE and UNICAP from Brazil and Universität Basel from Switzerland.
Our goal with this project was to assist the team to communicate their work and generate deliverables for different stakeholders, especially through data visualization. We developed a series of sketches until we reached two final formats: Radial and Network representations.
Radial Visualization
How could we make a chart more appealing? During our co-design process, we helped the research team to simplify and organize the categories more clearly and also decided to try to encode the numbers in a color palette.
Those decisions already made a big difference, but it wasn't enough to make the information stand out as we wanted. At this point, we found ourselves with the ideia of a Mandala, or a Radial Visualization, but it would only work if it made sense to the research team and if it was accurate to the result of the project.
At the end, what first seemed like a bold choice for a scientific visualization, turned out to be a very intuitive way to understand the correlations presented by the study. The result was an odds ratio visual comparison of 34 indicators in 5 different categories, a highlight session of Bolsa Família and a color blindness proof palette.
Network Visualization
Our goal was to visualize complexity. What combinations of Social, Health and Economic characteristics have the greatest impact on increasing or decreasing the risk of certain birth outcomes? How big is this impact? How many babies are affected? There were too many questions to be answered in just one image, which resulted in a total of 5 network visualizations.