Skip to content

Latest commit

 

History

History
50 lines (28 loc) · 4.15 KB

resources.md

File metadata and controls

50 lines (28 loc) · 4.15 KB

DATA 22700 Resources

Here are references and resources for the course separated by topic. This is where we will share links to readings, tutorials, examples, and programming tools.

Software

This is where we'll post links to software tools and demos, separated by topic.

Getting started with computational notebooks

Programming tools

  • Basic pandas tutor demonstrating common dataframe operations
  • Intro to pandas in Google Colab
  • The primary visualization API we teach in this course is Altair. Check out their user guide, example gallery, API, and more on the same website!
  • Altair is actually a wrapper around a JavaScript library called Vega-Lite. If you're interested, check out the Vega-Lite example gallery to compare its syntax with that of Altair.
  • From the creators of popular visualization toolkits like D3, Vega, and Vega-Lite, the University of Washington Interactive Data Lab has graciously made their visualization curriculum public! The notebooks posted here give an excellent walkthrough of some topics we'll cover in this course.

Demos

Readings

Here are optional readings for the course separated by topic.

The value of visualization

  • A well regarded academic article on the value of visualization.
  • Obligatory sharing of Tufte's Visual Display of Quantitative Information. Classes like this one often assign the first three chapters of this book as reading, probably because Tufte's work is rich with examples. However, Tufte sometimes asserts as design principles ideas that don't hold up when subjected to empirical scrutiny. His work is still viewed by the visualization community, with skepticism, as a wonderful resource.

The science of visualization design

  • This paper by Cleveland and McGill set a precedent for treating visualization effectiveness as an empirical question. Many papers have followed up on this work, and the findings remain roughly intact.
  • Mackinlay's APT paper lays out his expressiveness and effectiveness criteria for visualization design. This paper kicked off a long line of work on visualization recommender systems.
  • Particularly interesting violations of the expressiveness principle (a.k.a. "tell the truth and nothing but the truth") occur when people's expectations about what a certain kind of chart will show are violated. Among other sources, these expectations are informed by graphical conventions, such as the expectations that people have about the semantics of bars and lines addressed in this paper by Zacks and Tversky, which I mentioned in class.
  • A few years ago, some of my colleagues at Northwestern decided there was too much visualization research for practitioners to keep up with. They led an effort to write this review article summarizing the science of "what works" in data visualization design.