Printable Version of this PageHome PageRecent ChangesSearchSign In
I found the lecture to be extremely interesting and enlightening. Showing relationships in large data sets with many variables is a nontrivial task, and I was impressed with the ability of the software packages presented to accomplish that goal. Much of what was discussed reminded me of a book I read recently called Freakonomics, as the drawing of interesting conclusions by asking interesting questions from large data sets seems to be an underlying theme.

I suppose that in general I didn't fully appreciate the divide between traditional 3D data visualization for scientific data and the more generalized viewing of higher dimensions. I was a bit disappointed that there really isn't a good way to view data with more than four or five degrees of freedom at the same time. The overall technique matched my expectations, in that it was robust filtering of data constrained to low to mid level degrees of freedom. I am curious as to whether the software packages presented will allow you to switch quickly between completely different representations of the data to help find the view that provides the most clarity. I am curious how widespread utilization of tools like these to draw conclusions are in the sciences. A key issue in the presentation is the displaying of the uncertainty and I personally feel that the fading technique is the most intuitive and provides the most clarity.

Last modified 4 October 2007 at 5:33 pm by cobb