OK, in order! (And I haven't really thought about it yet; been...

OK, in order!
(And I haven't really thought about it yet; been out tonight.)
Genzr: Sure about the discontinuity but if you can model the expected price without the event.... (and probably with a time delay to let it settle down again). Volumes and ASX/NYSE would all be influences.
Toto: Briefly, regression analysis or better SEM can be used in situations like this. You can use these techniques to check how much of the price (changes) can be modelled by the knowable/measurable variables. If it's 90% it's a good model and you can check which variable's contributing how much. If it's 20% you can look at what's left over (the residuals) and how they (for example) correlate with 'judged' sentiment over time.
Copper: Correlation is of course important but often misused when there are multiple factors (especially if some aren't measured). I won't go into it...
Wllm: Thanks for that and I'll certainly save those data. (Date and Basket price is enough.) More data, especially over a bigger spread of time (eg the happy days on the way to $2.70) would be useful as you'd need over 50 data points (times) to consider fitting a half decent model.

I'll go quiet on this for a while because I have lots on my plate but will have a potter!