Some chart types are not scalable
Dissecting charts from a Big Data study

Update on Dataviz Workshop 3

My chart making workshop has passed the point where each participant (except one) has presented the first draft of his or her project, and the class has opined on these efforts. Previously, I posted the syllabus of the course here. Also catch up on previous updates (1, 2).

So far, I am very pleased with the results, and importantly, the students have given rave reviews. The in-class discussions have been very constructive, and civil. In every case, the chart designer went home with a few ideas for improvement. The types of issues that came up ranged widely. Here are some examples:

  • Figuring out what the message is in the data set
  • Thinking about what other data can be obtained to clarify the message
  • Discussing the level of detail appropriate for a legend
  • Dealing with data with a large number of small values
  • Because we have a color-blind student, we can examine how charts appear to the color-blind reader
  • How to reduce the complexity of a chart?

As the course draws to a close, several students have expressed an interest in keeping the class together via a meetup group or something similar. I'm thinking about how to accomplish this.

One lesson learned so far is that a few students got stuck trying to restructure the data, and were late submitting their work. I should stress that all submissions in the course are work in process, and maybe I should offer some data processing help during the course.


The next workshop will be offered in the summer.


PS. Don't miss Andrew Gelman's summary of his graphics tips here.




Feed You can follow this conversation by subscribing to the comment feed for this post.

The comments to this entry are closed.