I was hoping that it would be something to do with correlating google trends data with price and volume movement. This is cool, too -- it's pretty, and free, so don't take this particularly seriously. Currently, I believe I can get far more information on BitcoinWisdom. I'm not sure that TII shows me anything that certain KDJ parameters doesn't.
The real issue currently I'd have with using cointensity is that there is nothing I can tweak, so not only can I not customize it to my needs, I can't explore to see past TII response to Bitcoin market events to get a mental picture of what TII can and can't show me.
Constructive criticism, is all that I intend. Obviously, Cointensity is more than most BCT denizens are doing.
Thanks for the critique! I put this site together primarily for myself so I didn't need to update an Excel spreadsheet everyday. The charts I've included are simply those that I like to use to get an idea of how the bitcoin price is trending in the market. Anyone interested in doing their own detailed technical analysis should definitely use a site like BitcoinWisdom.
I like your idea of trying to correlate Google trends with the change in the TII over time. Are you thinking the historical TII (using all historical prices from Bitstamp) plotted against the trend in searches for "bitcoin"?
Since Big Data is one of my "things", I've spent a decent amount of shower time thinking about it. As such, if I were to find the time for such a project, what I'd look for falls into two broad categories to start. Let me preface by explaining my definitions and perspective a tad: Everything is correlated, somewhere in the range of 100% to 0%. IRL everything is in the middle. 100% is impossible unless it is literally the thing correlated with itself (like a mathematical identity); 0% is impossible because we're on earth and all humans, etc -- kinda like how there are no circles outside of math.
Cat 1: Things which either are obviously correlated or in which I'd expect to see strong correlation. Price, volume, trending terms like Bitcoin, cryptocurrency, satoshi, etc.
Cat 2: Things that are less obvious that I wouldn't expect to be significant, that turn out to be strongly correlated. These are fun to uncover using big data. Then you probably go about trying to understand WHY.
Then in both categories, you find the ones that are most indicative of things we might care about: Impending price moves, predicting bubbles, predicting crashes. There are definitely ways of using data like this to get interesting probabilities of things. It's the same math used to find new ways to predict weather, earthquakes, bird movement, disease propagation, etc. Google did a fun thing recently where they correctly predicted a local flu outbreak based on a seemingly insignificant uptick in flu queries in an area, correlated with weather patterns that indicated cold weather of the type that dampens immune function, and other stuff. Forget what journal I read about it in. You get the idea. Sky's the limit in 2014, my friend!