2010-05-11 class notes

[Jesse Cirimele]

Discussion:
what are people bad at doing?
-tracking more than one thing at a time (lose one digit's
multiplication when you try to do another)
-people have to use tricks to do mental math better (better visualization, etc)
-random makes it harder
-this weeks's reading built well on "the power of representation"
-we adapt to what we need to know (remembering cell phone numbers)
-you remember and you get good at what you are doing, we are used to
dealing with cartesian coordinates so we quickly understand those
visuals
-slide rule: use visual faculties to do math

what are people really good at that are hard to translate into computers?
-recognizing faces
-parsing sentences, not really visual so how might you extend?
-using context to figure out what something might mean
-backfilling (optical blind spot)
-we are good at foreground vs background
-making sense of malformed input ( a mangled sentence )

good and bad about challenger ( O-ring ) info-graphic:
good:
-lots of rockets (where is the damage happening, locality, about rockets)
-ordered in time (see trends)
-pulls people in ( people know what it's about and want to find out more )
bad:
-very cluttered ( hard to pick out individual rockets )
-suggests 48 rockets on space shuttle rather than 48 SRMs ( or o-rings )
how would you design it to be better:
-answer what you want to show? temp vs damage? manufacturer?
how would you show temp vs damage?
-use color
-scatter plot

it's not just what is the temperature, but what is the temperature on
the day you are going to launch?

even if you know what's "good" for visualizations, there may be a very
high cost to do it ( time, skills, tools )

motivation for second paper: people don't often work alone. there are
other people, annotations, interactivity.
-like that everyone has access to all of the information ( don't have
to just believe what you're told )

how complete is the analysis of the second paper?
-depends on user base. need critical mass of users to get past
"random" and into trends.

example from text: stock market crash with jumping trader. how do we
know if that annotation is moving it forward?
-any system that people can use for play as well as work is probably
valuable. play is a good indication that people care.

quick sense.us demo...