This week in class I decided to start working on parsing relationship-based email text from different periods of time. This is a screen-capture of the initial program running through processing. This pulls and counts instances of words from the same week in April of 2005, 2006, 2007 and 2008 respectively. As there are fewer instances of the words, the red gets darker. As there are more instances of the words, the red is lower in alpha value. The thinking was that it was the words that were typed less frequently that were actually more significant.
This is video of the applet in action:
Screen capture
I'd like to continue with it for the midterm week project, particularly getting more content together so the visualization can be more meaningful, and maybe having the program recognize specific words that could be pre-determined as either negative or positive, or both.. still working on the logistics of that.
posted by Amanda @ 10:24 AM,
