There are a few ways to do this. The simplest way is to create a blank list in your initialization, and append values to the list with each bar. You can also use deques if you are familiar with them.

def initialize(context):  
    context.security = sid(8554) #S&P 500  
    # create the blank list  
    context.past_prices = []

def handle_data(context, data):  
    # every bar (either minutely or daily) append the security's price to the list  
    context.past_prices.append(data[context.security].price)  
    # let's log the first bar's price  
    log.info(context.past_prices[0])  
    # now let's log all the prices we have stored  
    log.info(context.past_prices)  
    # you don't have to just store the price, if you want you can also  
    # store volume data or something like a moving average for each day  

Does that help?

With regard to moving averages you can take the price (or shorter moving average) and divided it by the longer moving average. If its greater than 1 the cross is a buy. If its less than 1 the cross is a sell. (This would mean buying not exactly when the two moving averages intersect but when the shorter actually crosses over the longer). As Gus said, there are also a number of other ways to tackle the issue of storing relevant data.

I tried creating a blank list but it appears quantopian doesnt recognize it. I get errors anytime i tried using the list. I declared the empty list both
as context.list = [] which doesn't work and gives me an error and as a global list which doesn't seem to work as well. Also the prices I would append are the minute interval prices. Is there a way to record the day closing price and create a list of daily bars if you will?

@Daniel, you can look at a recent thread as an example. In my backtest source code I used a deque to store the 5 most recent RSI indicators. You can use the same idea to store any value you want.

https://www.quantopian.com/posts/can-lamda-be-used-to-look-back#51e5f09bd3cf61daa0000139

from collections import deque  
import numpy as np

# initialize  
context.list = deque([], maxlen=15)

# append value (deque configured to keep 15 most recent)  
context.list.append(price)

# get average  
avg = np.mean(context.list)