Skip to content

Instantly share code, notes, and snippets.

@lxbarth

lxbarth/dataviznotes.md

Created November 21, 2011 04:08
Show Gist options
  • Save lxbarth/1381587 to your computer and use it in GitHub Desktop.
Save lxbarth/1381587 to your computer and use it in GitHub Desktop.
Download ZIP
Raw
dataviznotes.md

First: this is awesome work, the below obviously focuses on what's worth improving.

Audience

I wasn't 100 % sure from which angle to judge the paper:

  • What's the intended audience?

Related: maybe this paper should cut faster to the meat and don't spend so much time explaining a methodological framework of analyzing information viz tools? How would this read as a large blog post or CT article? I'm aware that thesis have an expected format. Still, could the paper get faster to its point?

Notes

  • In general, I'm missing geographical data vizualisation, mapping. I get that the paper focuses on graphing, but I'm missing at least the distinction.

  • Narrative Viz: http://one.org/us/actnow/horn.html is a great example

  • Curious: Why did you not include voronoi in network?

  • The Les Miserable matrix is awesome, did you do it?

  • I'm sometimes missing the browser support angle. It is critical in practical application and the fact that canvas is not supported in IE<=8 is a biggie.

  • Not sure I follow here: (3.4.5)

The implementation of a node forms a single instruction that operates on multiple data items. This means that visualizations can deal with any number of data items, which is particularly interesting in the context of Information Visualization.

  • I love Flot - it's a great micro library and maybe worth a mention. We use the crap out of it.

  • In practical terms, where do you see the sort of platform independence that uv.js offers come into play?

  • Is that true? SVG 'performs better' than HTML5 canvas as in 'faster'? Which browsers? 2008 is a little while ago. "They found that SVG performs better than HTML5 Canvas, given that the number of objects do not become too high."

  • This is contradicting the benchmarks (emphasis added):

While D3’s expressive syntax and deep integration with developer tools is suitable for many visualization tasks, Unveil.js may be a good fit when it comes to high performance rendering involving a high number of animated objects (such as particle systems) or approaching complex visualization tasks through object-oriented abstraction

  • I love the comparison, not sure how i feel about the scoring especially as the author's sofware wins :) I'd just leave it out and add a comparison table and use symbols instead of numbers (full dot, empty dot, dash, like when you buy a car?).

Minor:

  • There's a "constrast"
  • "HTML5 is ahead" should be "HTML5 Canvas is ahead"

Unclear:

This principle is referred to as The Information Seeking Mantra and serves as a starting point for creating visualizations. Based on that, Shneiderman also proposes a Type by Task Taxonomy (TTT) of Information Visualizations. He assumes that users are viewing collections of items, each having multiple attributes.

Based on Shneiderman’s work, the following sections introduce Tasks and Data Types, complete with illustrative examples

The jump from "TTT" to "Tasks and Data Types" is not clear - how do these terminologies connect?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment