Data Visualization – the Why and How

Sho: I recently gave a talk on data visualization at the International Conference on Infant Studies (you can find my slides, along with the other wonderful talks on power, preregistration, and ethical data peeking here on the OSF). I also played the German Cats and Dogs Scientist in the barbarplots campaign on better data visualization (on the same topic, Article 1 and Article 2 on why bar and line plots hide differences in underlying distributions). In fact, being part of the barbarplots team was my entry point into thinking more about the importance of visualizing your data in a maximally informative and honest way. Informative means finding a good balance between simplifying/summarizing and showing the underlying data structure. Honest means not (accidentally) hiding important aspects of your data. Mahiko is my office mate and – this is something I discovered while preparing the above talk – an enthusiastic data visualizer. That’s why I asked him to put together our (mostly, his) favorite data viz resources.

Mahiko: I don’t remember when I initially got interested in data visualization, but I’m pretty sure one thing in particular attracted me: the beauty of being able to combine aesthetically pleasing designs with intuitive information. “Visualization is communication”, says John Rauser in the below-linked “amazing youtube lecture” (it really is): by creating a visualization, we are trying to transfer a piece of information from our minds (or our data), to another person’s mind, as quickly and efficiently as possible. Making it intuitive helps with the speed of this transfer, and making it beautiful attracts the interest of the receiver, ultimately helping with his/her’s understanding.

In a world in which the speed and amount of new information to process is quickly exceeding human capacity (~20 still unread articles in my open mobile chrome tabs, I’m looking at you), I believe that good visualizations are not a luxury anymore, but a duty of scientists.

So here you go – our assembly of data viz resources. If you have more we’d be happy to add them!

 

Design principles

Old but gold paper on how humans see and understand data visualizations. If you prefer a more modern format, an amazing youtube lecture on the same topic.

Incredible guide on color palettes for sequential and categorical data, also includes colorblind palettes.

 

Plots and data types

Comprehensive and fantastic guide on which plot type to use depending on your data. Also has Python and R code.

Top 50 R ggplot2 plot types for different types of data with code.

Suggestions on visualizing ERP data (also applicable for time course data in general)

On why pirate plots are a good solution for distributional data because they combine Raw (data), Description and Inference.

 

Comprehensive resources

Online tutorial on communicating your data from color scheme to plot type

Online book 1 and online book 2 on data visualization including helpful R code.

 

Random inspirations

Blog post with python code to generate line plots and other FiveThirtyEight style plots

Beautiful visualizations on Reddit’s Data is Beautiful

Ever wondered whether pop lyrics are getting more repetitive? The best visualizations we’ve ever seen on random topics.