FYI: I've come across this challenge that would probably lend itself well to a NuPIC/HTM based approach.
https://www.innocentive.com/ar/challenge/9933692 The challenge deadline is March 9. Unfortunately, I probably don't have the necessary experience with NuPIC to field a credible entry by then, but I would be willing to help out if anyone here is interested in pursuing it. If nothing else, it looks like they are using extremely large, time-series data sets that might be an interesting test case for the wider NuPIC development effort. Regards, Eric M. Collins
