Here’s a Neoformix analysis of Twitter conversations that happened yesterday and today about Obama’s running mate.
If I understand correctly, Neoformix only analyzes a set quantity of Tweets (I don’t have the specifics.) Knowing that this number is the constant, we have something to analyze everything else by.
For example, note how:
- More tweets are made about Bayh (higher number on y-axis) less frequently (longer period on x-axis.) Conversely, less tweets were made about Biden more frequently.
- Bayh’s Tweets “exploded” during the early part of the afternoon (this was tied to yesterday evening’s bumper sticker debacle) then remained small (when Obama announced Biden’s pick,) while Biden’s Tweets was relatively stable throughout the night.
- Bayh’s Tweets often mentioned the other two potential candidates (Biden and Kaine,) while Biden’s mention his first name and Obama’s—with bits of McCain here and there.
My theory: in predicting political candidate, a Twitter stream’s relative stability is a much bigger factor than its quantity.
But this got me thinking. Could this also mean that the factors lined out above can be used to predict who McCain’s running mate will be?
I’ll leave the thinking part to you.