1. Read in the Schlumberger stock data, put the Date column into the index, and convert it to a DatetimeIndex:

>>> slb = pd.read_csv('data/slb_stock.csv', index_col='Date', 
                     parse_dates=['Date'])
>>> slb.head()

2. Select the closing price as a Series and use the describe method to return summary statistics as a Series:

>>> slb_close = slb['Close']
>>> slb_summary = slb_close.describe(percentiles=[.1, .9])
>>> slb_summary
count    1895.000000
mean       79.121905
std        11.767802
min        51.750000
10%        64.892000
50%        78.000000
90%        93.248000
max       117.950000
Name: Close, dtype: float64

3. Using boolean selection, select all closing prices in the upper or lower tenth percentile:

>>> upper_10 = slb_summary.loc['90%']
>>> lower_10 = slb_summary.loc['10%']
>>> criteria = (slb_close < lower_10) | (slb_close > upper_10)
>>> slb_top_bottom_10 = slb_close[criteria]

4. Plot the resulting filtered Series in light gray on top of all closing prices in black. Use the matplotlib library to draw horizontal lines at the tenth and ninetieth percentiles:

>>> slb_close.plot(color='black', figsize=(12,6))
>>> slb_top_bottom_10.plot(marker='o', style=' ',
                           ms=4, color='lightgray')

>>> xmin = criteria.index[0]
>>> xmax = criteria.index[-1]
>>> plt.hlines(y=[lower_10, upper_10], xmin=xmin,
               xmax=xmax, color='black')