Recently I was invited to give a talk to two cohorts of Strategic Data Project fellows. I was asked to speak about using data visualization to help inform decision-making of policy makers. At the same time, the group had a lot of variation in their interest and prior experience with data visualization. For my talk I decided to try to fit a little bit of everything into a 60 minute discussion:
- Principles and why they matter
- Best practices and routines to follow
- Branding and when style is allowed to trump function
- Visualizing very large datasets
The talk is very specific around education data, but education data has a lot in common with datasets from any number of other fields, and so the talk might be useful to others interested in learning more about data visualization, education data, policy making, and/or all of the above!
Additionally, this was my first attempt at making a slideshow using the slidify package for R. My previous workflow involved using knitr and pandoc to make slideshows, for example, for the R Bootcamp. However, I wanted to see if slidify provided an upgrade. This package has many fantastic features that I loved, including the ability to publish the slides directly to GitHub, which is how I distributed them. This is great, but for projects I don’t want in public, distributing the slides offline is a little less straightforward.
The big advantage for me with slidify is the style flexibility. The pandoc HTML5 slides are a little vanilla in their style and somehow seem more difficult to customize. slidify seems to have found a way to make this much more direct, which I appreciate.
I think learning both has been helpful. I hope to post more soon on pieces of CSS and HTML that make web-based slides superior to their PowerPoint ancestors, and streamlines the styling process.