On predictive analytics

We want to predict things. We want to know the future if we can. So what are some examples of that? Okay so three M sales Peter's team. So three M would like to know what factors influence the sales of Peter's team. For instance, if we cut the price, how much more will we sell? If we advertise how much more will we sell? Is the season of the year to what we sell? Is a holiday important to what we sell? Do we sell more on Memorial Day? Do we sell more on 4th of July? And that can help us increase the profitability from the three M tape. Another example near and dear to your heart if you're a sport fan is football point spreads.

Okay, or basketball point spreads. How do the odds makers in Las Vegas come up with the point spreads on sporting events? So, we are taping this video during the NBA finals in 2013. And basically in gain for, the spurs at home were a two point favor. And they got creamed by Miami. And that engage, so the two point favor by the Spurs means, the bookmakers essentially think the most likely outcome is the Spurs win by two. Well, that's basically half the money will go on each side, if you set that point Spur. Then, in game five They flipped it til Miami was favored by half a point.

Okay, so that means they factored in what happened during that game four. And they really change, changed the view of what would happen and made the Heat the favor, instead of the Spurs based on the Heat winning game four by a lot. What are some other examples predicted analytics? Well. We've unfortunately probably known one of our relatives or our friends, who's gotten very sick, and you would like to know if your friend has cancer, or your relative has cancer. Can we predict the chance of survival? Okay, what treatment, course of treatment given a disease, will give us the best chance for survival? Of course cost must enter in to that.

So, what factors do we need to know to predict the chance of survival? Many of you are probably familiar with Nate Silver who rocketed to fame during the 2012. Presidential election. He picked all 50 states correctly how they would go Romney verse Obama and so basically, what factors did he use to make those predictions? He also makes predictions on the World Cup. He makes predictions on the Oscar's. I don't think the Oscar predictions went so well. But basically those are some examples of where we want to use data to make predictions and that's what predictive analytics is all about.