yesterdays_close is not a number, it's a pandas dataframe. You need to index into it if you want get the exact number you care about. Hooray dynamic typing.

Thank you, Simon. I figured out how to index the DataFrame using this snippet of code close_index = yesterdays_close.iloc[0][context.stock]
That was the answer I needed and here's my code.

from pytz import timezone  
from datetime import datetime, timedelta  
from zipline.utils.tradingcalendar import get_early_closes  
from quantopian.pipeline.data.builtin import USEquityPricing  
import pandas as pd

def initialize(context):

    set_commission(commission.PerTrade(cost=0))  


    set_slippage(slippage.FixedSlippage(spread=0.00))  
    context.minutes_after = 5  
    context.stock = sid(40634)

def handle_data(context, data):  
    yesterdays_close = history(1, "1d", "high")  
    log.info(yesterdays_close)  
    close_index = yesterdays_close.iloc[0][context.stock]  
    todays_high = data[context.stock].open_price  
    log.info("current price: %f" % todays_high)  
    pct_change = (todays_high - close_index) / todays_high * 100  
    log.info(pct_change)  
    cash = context.portfolio.cash  

    for stock in data:  
        stock_price = data[stock].price  
        plausible_investment = cash  
        stop_price = stock_price - (stock_price * 0.005)  
        buy_amount = int(plausible_investment / stock_price)  
        share_amount = context.portfolio.positions[stock].amount  
        capital_used = context.portfolio.capital_used  
        if capital_used == 0:  
            order(stock, buy_amount, style=LimitOrder(stop_price))  
            print "ordering shares"  
        if pct_change <= -2.0:  
            if context.portfolio.positions_value > 0:  
                order(stock, - share_amount, style=LimitOrder(stop_price))  
                print "selling shares"  
        if stock_price <= 1.90:  
            if share_amount > 1:  
                order(stock, - share_amount, style=LimitOrder(stop_price))  
                print "selling shares"


def market_open(context, data):  
    # Checks if it's market open at 9AM EST  
    loc_dt = get_datetime().astimezone(timezone('US/Eastern'))  
    # Backtesting starts the time at 9:31AM  
    if loc_dt.hour == 9 and loc_dt.minute == 31:  
        log.debug(get_datetime())  
        for stock in data:  
            context.daily_stocks[stock] = data[stock].price  
    else:  
        pass  
def market_after_open(context, data):  
    # Checks if it's market open at 9AM EST + however minutes you want to set  
    loc_dt = get_datetime().astimezone(timezone('US/Eastern'))  
    if loc_dt.hour == 9 and loc_dt.minute == 30 + context.minutes_after:  
        log.debug(get_datetime())  

Okay, double-check you are getting the day you want though. I think history always returns "today" as a partial bar at the end of the data frame, so to get yesterday, you might need to request two bars and then get the first one (at index 0).

Simon - yesterdays_close is actually just a placeholder, I should have renamed it to what I was actually trying to accomplish, which is getting the high price of the stock for the current day.
After the highest price is received I compare that to the current price. If the price below the highest price is less than or equal to 2% I automatically sell the stock.
There is also a function to sell the stock if it reaches or goes below the price at which it was bought.
This lets my position roll while still maintaining relative control over volatility.