Interesting, if flawed, article about using 'big data' for indie film development and promotion.
I'd point to an article I posted earlier in the week about IBM's SSI experiments in Bollywood as a possible signpost to the future.
Large data sets are no longer unwieldy, or unobtainable. The clever part is knowing a) what metrics to apply and b) how to appropriately act on the findings.
Building the tools are easy. There's even a course on Coursera about Machine Learning. What is really tough is figuring out what metrics to apply. Machine Learning is about take values x and y and being able to predict z. Give me the x and y and I'll build to tools.
And, once you've discovered the 'z'… what do you do then? In the age of privacy, knowing patterns is usually only helpful when you can identify types of people, not individuals.
There is already a great tool available that gives you a peak into the biggest user data collection on the planet – and lets you market to it. It's called Facebook Ads. With Facebook Ads you can say 'show this ad to people who are, or might be, interested in xyz'. It's an incredible powerful advertising platform, greatly under-utilised (like almost everything!) by indie filmmakers, mostly due to them not knowing its there.
One day I will write a book 'Internet Marketing Principles for the Indie Filmmaker'.
Anyway…read the article. He's a Raindance Masters mentor.