Predicting Campaign Response, Figuratively and Literally (and Burritos)

I am always interested in seeing new applications, or more extensive applications of predictive modeling. A lot of times, we use the figurative term “campaign response” to talk about predicting generalized response behavior. However, we’re in the middle of a season where predicting campaign response is far more valuable when you’re speaking literally. Oh, and for the bit about burritos, see the bottom of the post.

One article in particular jumped out as I was browsing around, and I point to it primarily because it is representative of predictive modeling in politics. Apparently, Ted Cruz’s campaign has been taking full advantage of the insights that data can reveal. It would seem it is paying dividends, as well, if you are up to date on the latest polls. There’s certainly still a gap, but there has been a very real turning point for Cruz’s standings in the field.

As you can see, around December 2015, Cruz's poll numbers took a distinctive change for the better.

As you can see, around December 2015, Cruz’s poll numbers took a distinctive change for the better.


And as we’ve mentioned previously, although Predictive Analytics (for any use-case) is getting much more attention lately, it isn’t some fly-by-night solution. Rapid Insight software was used in a case very similar to this: former Senator Lieberman’s campaign.

So are there fields you wouldn’t expect predictive modeling to be useful in? After all, it’s nearly taken over politics at this point, it’s used in sports, higher education, retail, weather, and probably the most important use of all, predicting burrito-deliciousness.