def bollinger_bands(data, n, std_mult=2.0):
catch_errors.check_for_period_error(data, n)
sma = simple_moving_average(data, n)
smstd = sma.rolling(n, min_periods=n).std()
upper_bb = sma + std_mult * smstd
lower_bb = sma - std_mult * smstd
return lower_bb, sma, upper_bb
def momentum(data, n):
"""
Calculate the Momentum af the close price of a stock over n
:param data: pandas.DataFrame
:param n:
:return: pandas.DataFrame
"""
catch_errors.check_for_period_error(data, n)
Momentum = pd.Series(data['Adj Close'].diff(n), name='Momentum_' + str(n))
return Momentum
def simple_moving_average(data, n):
"""Calculate the moving average for the given data.
:param data: pandas.DataFrame of prices data
:param n: int number of periods to look back
:return: pandas.Series of moving averages
"""
catch_errors.check_for_period_error(data, n)
MA = pd.Series(data['Adj Close'].rolling(n, min_periods=n).mean(),
name='SMA_' + str(n))
return MA
def exponential_moving_average(data, n):
"""
Calculate the expenential moving averge af the close price of a stock over n
:param data: pandas.DataFrame
:param n: number of periods to look back
:return: pandas.Series of exponential moving average
"""
catch_errors.check_for_period_error(data, n)
EMA = pd.Series(df['Adj Close'].ewm(span=n, min_periods=n).mean(),
name='EMA_' + str(n))
return EMA
def stochastic_oscillator(data, n):
catch_errors.check_for_period_error(data, n)
percent_k = [
100 *
(data['Adj Close'] - data['Adj Close'].rolling(n, min_periods=n).min())
/ (data['Adj Close'].rolling(n, min_periods=n).max() -
data['Adj Close'].rolling(n, min_periods=n).min())
]
return percent_k
def price_SMA_ratio(data, n):
"""
Calculate the price over SMA ratio
:param data: pandas.DataFrame of prices data
:param n: int number of periods to look back
:return: pandas.Series
"""
catch_errors.check_for_period_error(data, n)
SMA = simple_moving_average(data, n)
ratio = data['Adj Close'] / SMA
return ratio
def bandwidth(data, n, std_mult=2.0):
"""
Bandwidth.
Formula:
bw = u_bb - l_bb / m_bb
"""
catch_errors.check_for_period_error(data, n)
lower_bb, mid_bb, upper_bb = bollinger_bands(data, n, std_mult=std_mult)
bandwidth = 100 * (upper_bb - lower_bb) / mid_bb
return bandwidth
def relative_strength_index(data, lookback):
"""
Calculate the relative strength index
:param data: pandas.DataFrame of prices data
:param lookback: int number of periods to look back
:return: pandas.Series
"""
catch_errors.check_for_period_error(data, lookback)
daily_rets = daily_returns(data)
rsi_list = [np.nan] * lookback
for day in range(lookback, data.shape[0]):
up_gain = daily_rets.iloc[day - lookback + 1:day +
1].where(daily_rets >= 0).sum()
down_loss = -1 * daily_rets.iloc[day - lookback + 1:day +
1].where(daily_rets < 0).sum()
if down_loss == 0:
rsi_list.append(100)
else:
rs = (up_gain / lookback) / (down_loss / lookback)
rsi_list.append(100 - (100 / (1 + rs)))
rsi = pd.Series(rsi_list, index=data.index)
return rsi