Comment by Pinckney | Hacker News Reader

Comment by Pinckney | Hacker News Reader

This is a beautiful visualization, but it is very hard to parse, and I have no idea as to what problem it was designed to solve.

With regards to this visualization in particular, it is hard to judge which majors are disproportionately represented in any individual field.

There are in some places gaps in the spacing which make it appear as if a particular area is underrepresented; for instance, observe the apparent gap in Div I majors between Engineering and Health/Medicine.

It is very hard to judge the relative distribution of fields for small majors, such as Geoscience.

Path density seems exaggerated when the paths are drawn between nearby cells; compare the apparent density of paths between Div I <-> Arts/Writing/Social Science/Govt/etc with Div I <-> Education/Health. This seems to be because of the thinning of lines in the middle of the figure.

In reading about this type of figure, I managed to find possibly the most incomprehensible infographic I've ever seen:

http://circos.ca/intro/general_data/img/circos-car-purchase-...

I have a feeling that for the coming year or so, these visualizations will be overused in much the same way animated web site intros were in the late 90's. Because right now they are still new enough that the casual viewer will inevitably be impressed by their pretty colors, even if they do not particularly clarify any information. I suspect a slight backlash by late 2013, then a convergence towards legitimate use thereafter.

They're already overused in genome biology. They do a beautiful job of representing interchromosomal rearrangements in cancer, but don't actually convey a whole lot of information.

Yes! I hate when people have a figure of some crazy bioinformatics they did that doesn't actually convey any information to me as a reader.

It could have been a wordle.

Just imagine.

It could have been a wordle.

If you mouseover the specific visualizations on the righthand side, it will "filter" the main one and show only one major group.

That isn't much better though. With very few exceptions, it looks like regardless of major people tend to enter various fields in very similar proportions. If the data were displayed in a table, I have a feeling the numbers would tell a much different story.

Reply on news.ycombinator.com