It has been… a few years actually, and I finally have some ideas to share on my blog. Indeed, I will make an effort to put out content on a more regular basis than once every few years. The reason is simple: with family life taking priority, I travel a lot less these days and social networks are no longer what they used to be. So, this is my new-yet-old approach to share with the community some of tips, tricks, and other ideas that I come across whilst “sciencing” on a daily basis.
Today, I only share a small but cute idea for visualizations. To make data pretty, it is important to carefully choose the colours. It turns out that one fruitful approach is to take a set of colours from a famous painting. The original artist often composed the work with a certain aesthetics and had actually put some time and effort into the choice of palette. Blake Mills applied this idea and generated colour palettes from works at the Metropolitan Museum of Art (New York). You can see for yourself on his Github repository whether there is anything to your liking.
I know nowadays any self-respecting plotting library (e.g., Matplotlib, Seaborn) and the famous Colorbrewer website make available well-calibrated colour schemes in a single line of code or a few mouse clicks. But that also means if you want to stand out alternatives can be interesting.