I would do a disservice to the Adobe and University of Washington team that worked on the Zoetrope Project if I attempted to explain how this works. The video will do a much better job than I could hope to do.
The amazing thing that this tries to do is pull time sensitive data that could be unrelated into a more relational context. You will see an example in the video of a user that sees a spike in gas prices. By researching the date further the user notices a drop in price of oil on another website and yet third website highlights a story where Congress voted on a spending bill for Iraq.
For those that know what they are doing I think this will be a great tool. The danger is in making inferences where a relationship does not exist. Imagine if someone suggested that the price of gas spiked in the example above because of a solar flare. Tracking this information would be difficult as you could falsely conclude just about any cause/effect relationship you want.
This leads me to think about a site like Farecast and how they might leverage something like this to build an even more accurate system to advise users on airfare cost. How will something like this affect sales of market driven products and services? While this is still a pretty far out concept the day is coming where the relationship between price and other data are brought together. I am not certain the average user will use this, but rather the gatekeepers who will disseminate to the social networks.