Why don't you start with the closes first? Also, just how far back in history are you intending to go?
At closes, MMs tend to widen their bid-ask spread, so close bids and close asks are unusable. I have a few strategies that look good using last trade prices or mid-prices, but then when I tried on real bid prices and ask prices, the results were disappointing...
I use the Bloomberg notation and build up what I call a bid-ask-last table that is incrementally updated. I get the data as bloomberg sends me.
Well, good luck with all this, is all I can tell you... My point is that, before you can tackle the complexities of processing live data, you'd better have a good idea of how to organise it and store it. I suggest that you start with close data as a way of "dividing and conquering" the task.