overview
We’re using several data sources of Covid related data to try different views that might be helpful for public health and government officials.
You can find some of the original source files for the hyperglyph designs on github.
quick insight using a large scale range
Glyphs can hold dozens of data parameters, so we work to include as much information as is needed to really grok the complexity of a particular scene. But that can risk producing a complete, unintelligible mess.
So we create designs that show a few, key variables, giving them a large scale range so that differences between them can more easily be seen at a distance. In this example, it’s clear that the red spheres and maroon cubes have increased drastically in size over the last few months (each ring represents a months worth of data).
The clear visual indication of the change in significant indicators shows the user where to dive deeper into the visualization.
View of the drastic increase in cases and deaths in the recent months for North and South Dakota.
Details of South Dakota’s containment and closure policies around the time of the increasing cases and deaths.
zooming in for detail
The (very) small cylinders on each day’s rod don’t clutter the view when you initially look at the visualization. But because the data is present in the scene, it’s easy to quickly zoom in and start to understand some history and context to the phenomena of the dramatic increase in cases and deaths.
Lighter colors show categories where the state has no policy. Darker colors show categories where the state has strict policy. Zooming in, it’s easy to see the state’s approach to government direction for closures and containment.
It’s long been a goal of ours to work with people who develop ML models. There should be a few different applications of glyphs to the process of developing more accurate models, and with a recent data analytics challenge, we found a way to test a few of them.