The October meeting of the San Francisco Bay Area R User Group held at Santa Clara University consisted of socializing, an intro, and three speakers. In the intro, host representative Sanjiv Das highlighted the curriculum and advisory board of the school's new MS in Business Analytics program. The first speaker, yours truly, reenacted Sara Silverstein's Benford's Law post using R and insurance industry data (see previous posts in this blog). In light of the yahoo email scandal that broke that same day, it was posed to attendees whether a similar "law" might be found to discriminate between harmless and harmful emails without regard to message content. The last comment from the audience seemed to capture the evening's temperament: "Snooping is snooping!"
The other two timely talks dealt with election forecasting.
Mac Roach previewed a new online app from Alteryx to predict U.S. election results at the neighborhood level. Equally interesting was Mac's countrywide display, which was the first time I had seen graphical evidence of the increasing polarity of the American electorate, a disturbing trend IMO.
The last speaker, Pete Mohanty, spoke about presidential forecasting using bigKRLS. I was struck by the existence of a closed form solution to the problem. Pete's slides can be found here.
For a brief summary of the meeting, see BARUG's Meetup site.
No comments:
Post a Comment