This article, Subtleties of Color by Robert Simmon was incredibly insightful and full of information. I do not have much critiques here as it was more information driven than an opinionated piece; but I will go into details about the lessons I learned and what I wish to incorporate in the future of making data visualizations.
I enjoyed learning about how differently the human brain perceives colors compared to a computer (linear versus nonlinear, symmetrical versus uneven, differences in sensitivity to lightness and certain hues) and I found it super useful to get a (very) basic breakdown of color theory: what is saturation, lightness, and hue.
Diving into data types and the creation of color palettes to represent the differing data types was interesting. Sequential datasets benefit from a color palette that primary changes in lightness (because people perceive lightness most easily), coupled with a difference in hue or saturation. Qualitative data would benefit from color palette with colors as distinct from each other as possible; it is best not to exceed 12 categories, but if you must, try to group classes together and then make sure of color saturation. Divergent colors benefit from having two distinct color hues with a neutral middle part, and the hues differ in saturation and lightness. I have worked with all the three data types he mentioned (divergent, sequential, and categorical) and I somewhat choice similar palettes to what he describes as good practice, but I had no idea why those were the ones I chose at the time. I did not take information-based decisions but worked from intuition, and I was fascinated to learn about why certain palettes work better than others and convey the information more accessibly. Speaking of accessibility, another big lesson I learned is how to choose palettes taking into account color blindness. That is a very valuable take away. Also, speaking of intuition, I really appreciate the section where he covers intuitive color decisions when it comes to eligibility (when working with multiple data layers, use muted colors)and color associations (blue for sky, green for vegetation) and points out that there are cultural differences in that aspect as well (blue for abstract, orange-red for malevolent).
I enjoyed the articles because it was information driven, and coupled with an emphasis for common sense and the subjectivity of human nature. This is essentially the best way to teach about color usage.