In the beginning of the COVID-19 pandemic, esp. compared to Italy, officials in Austria struggled to provide open data and analysis on the emerging pandemic. A lot of comminuty efforts popped up to collect and clean up available data. I documented the issues in an article on the citizen journalism blog semiosis.at.
I created a set of visualizations to keep track of COVID-19 in Austria based on data gathered by the community. All the code to transform the source data and create the visualizations is open source. Got it automated now so all charts can be updated daily. I use a workflow with Jupyter Notebooks + Altair/VEGA to develop the charts, then rollup/babel for optimization of the JS code (including IE11 compatibility, yay, that was tricky!). You can find all the charts here: https://walterra.github.io/covid-19/
I developed these examples to demonstrate how curated reports like this could benefit the public and deciders in contrast to more exploratory dashboards. These dashboards are useful for domain experts, but often they lack focus and explainability. I use the whole setup to teach best practices in data visualization design and workflows in the Visual Analytics for Data Science course I’m giving at Kufstein Applied University.
I liked the outcome of these charts in particular: Small multiples with each item focusing on one entity with the rest for context as gray lines in the background. It’s basically how scatterplot matrices are often done, here it’s line charts instead. I was lucky that Austria has exactly 9 counties to fit into that layout ;-).
Update July 2024:
Observable Framework formalises the principles I tried to lay out with the approach describe above (develop in an expert environment, then publish a curated outcome as a static website) in a much better way now! Wish it was available back then to teach data visualiation! It’s definitely worth diving into their blog: https://observablehq.com/blog
Comments powered by Talkyard.