Note: This was done for a class assignment, and had to stay within certain restrictions. If time permits, I plan to redo this project without the restrictions.

User Guide

To interact with this visualization, slowly mouse over the countries. The country will change color, and a tooltip will pop up with the name of the country and totals for that country. Additional information (in a bar chart) will also pop up, beneath the tool tip. The bar charts are normalized, so the number at the end of the “prematurity” bar indicates the number of prematurity cases for that country. The tool tip and bars change as the mouse moves over different countries.

Please note: the mouse must be moved slowly. If the interaction fails, move your mouse to another area of the map. There is a glitch with the text being selectable, even when not visible, and this can interfere with mousing over the countries.

Dataset

To obtain this dataset, I downloaded csv files from the World Health Organization (WHO) website. It contains global child mortality counts, per country and for various causes.

To get the original dataset ready for the visualization code, I needed to do a fair amount of preprocessing in R. Initially, I simplified the dataset, and dropped columns for various codes used by WHO. Records with missing data were filtered out and the country codes in the WHO dataset had to be translated to the codes used by TopoJSON. Additionally, I reshaped the data format so that it was more compact and easier to use in the visualization code.

Some of the columns are descriptive of the country, such as country name, codes and region. These have string values. There are also columns for the age group and year of the data sample. In this version, I selected a single age group and year from the full data set, and these columns have repetitive values. The rest of the columns are named by causes, and contain numerical data on the number of children who perished in that country from each cause. These values range from zero to over a million.

Why did this dataset? While this dataset is uncomfortable to view and work with, I thought I might find patterns that can be changed. If we see how these children perished, perhaps we as a global community can work together to prevent these conditions from continuing.

Data/Code Sources and References:

WHO data:

http://apps.who.int/gho/data/node.main.GBDC-by-age-group

TopoJSON:

https://github.com/mbostock/topojson

Technique(s)

The main visualization is a choropleth, of the values for “All Causes” of each country. The values were encoded on a color scale, with darker colors indicating larger values. Rather than a continuous scale, values were assigned to levels based on which range they fell into.

The mouse over detail shows bar charts, for that country’s data, with the width of the bars indicating the normalized values of each cause. Since the bar charts are normalized, the value of the “premature” cause is labeled with that country’s count. To avoid cluttering the visualization, only that bar has the number label.

Together, these techniques give overview + detail. The map gives the global overview, and the bar charts give details.

A note about design and color choices: I intentially designed a simple visualization and kept the color scheme calm. The content of the data set can be emotionally difficult to think about, so I wanted to make the visualization as unchallenging and relaxing as I could while also maitaining a simple and appealing elagant design. I had many ideas of how to make a dramatic statement, like using piles of toys to encode the statistics. However, I decided that putting the viewer at ease would result in more time spent really looking at the data, and possibly to possitive action.

What I Learned:

I was surprised to see that the leading cause was prematurity, and that this held true across almost all countries. Another leading cause involved birth complications and traumas. Together, these causes indicate that if we improve prenatal care and access to high quality midwifery, we could have a huge impact on child survival in infancy worldwide.

Other patterns from the data are also visible. Certain causes tend to cluster together in different countries. Group 1 (communicable disease and parasitic infection) tends to be present along with Sepsis and congenital anomalies.

How we can change the world from here:

This data visualization shows common dangers to infants around the world. If it can be used to emphasize the need for better prenatal care and birth support, here and around the world, we can save many children from perishing.

Profile:

Kristin Henry khenry1@dons.usfca.edu

Before studying as a Graduate student in Computer Science at USF, Kristin Henry consulted in web development, specializing in Science and Data visualizations. She also founded a non-profit dedicated to science literacy, and is an active generative artist.