Predictive analytics is this super-complex field that only statisticians and data scientists can understand, right? Well, perhaps it takes some training to do it well but it only takes two sentences to understand what it’s all about: In predictive analytics we determine the likelihood of something by looking at data about it. We do this simply by looking for similarities between that data and data from past cases where we actually know the outcome. So for example, the Not Hot Dog app, based on HBO’s show Silicon Valley, doesn’t actually know that it’s looking at a hot dog, it is simply predicting the likelihood that what it sees is a hot dog based on actual hot dogs it’s been shown (i.e. trained with) in the past. It doesn’t tell you that there’s only a 72% probability that what its seeing right now is a hot dog but, in fact, that is what it “knows.” Simple.
