def func(t):
data = read_ETF(t)
if data.index[0] < start:
data = data[start:end]
a, b = get_alpha_beta(data.Close, ret_type='simple', dspl=False)
df.at[t, 'alpha'] = a
df.at[t, 'beta'] = b
def show_dividend(ticker, start=None, end=None):
"""
Show annual dividend yield
input: ticker ticker of the stock/ETF
start start date, default '2000-01-01'
end end date, default current
Return a DataFrame
"""
from invest.useful import convert_time
from invest.get_data import read_ETF
from invest.calculation import add_dividend, get_dividend_yield
start, end = convert_time(start, end)
data = read_ETF(ticker)[start:end]
add_dividend(data, price='Close', adj='Adj Close', out='Div')
div = get_dividend_yield(data, price='Close', div='Div', style='simple')
plt.figure(figsize=(14, 4))
plt.bar(div.index, div * 100)
ret = data.Div[data.Div > 0].tail(6).to_frame()
ret.index = ret.index.date
print(ret)
plt.show()
return ret.T
def investment_performance(ticker, price, quantity, start, end=None):
"""
Calculate performance of given stock/ETF.
input: ticker ticker of the stock/ETF
start start date
end end date, default current
"""
from invest.useful import convert_time
from invest.get_data import read_ETF
from invest.plot import plot_day_price
from invest.calculation import add_dividend, get_return_vol
from time_series.functions import resample
start, end = convert_time(start, end)
data = read_ETF(ticker)[start:end]
plot_day_price(data)
plt.plot([start], [price], 'ro')
plt.show()
add_dividend(data, price='Close', adj='Adj Close', out='Dividend')
temp = data[['Dividend']][data.Dividend != 0].copy()
temp.index = temp.index.date
display(temp.T)
weekly = resample(data, style='week', method='close')
rv = get_return_vol(weekly[['Close', 'Adj Close']],
scale=52,
ret=False,
plotit=False)
rv['Total Return'] = data[['Close', 'Adj Close']].iloc[-1, :] / data[
['Close', 'Adj Close']].iloc[0, :] - 1
rv = rv * 100
rv['Gain'] = np.round(price * quantity * rv['Total Return'] / 100, 2)
display(rv)
print("Actual gain without reinvest: {:.2f}".format(
(data.Close[-1] - price) * quantity))
print("Dividend gain: {:.2f}".format(data.Dividend.sum() * quantity))
def show_yield(ticker, start=None, end=None, weeks=52):
"""
Calculate annualized return. Simple return is calculated
input: ticker ticker of the stock/ETF
start start date, default '2000-01-01'
end end date, default current
weeks number of weeks for each calculation, default 52
Return a DataFrame with three rows: Adj Close, Dividend, Close
"""
from invest.useful import convert_time
from invest.get_data import read_ETF
from invest.calculation import add_dividend, get_returns
from time_series.functions import resample
start, end = convert_time(start, end)
data = read_ETF(ticker)[start:end]
add_dividend(data, price='Close', adj='Adj Close', out='Dividend')
data['Dividend'] = np.cumsum(data.Dividend)
weekly = resample(data, style='week', method='close')
weekly = weekly[(weekly.shape[0] - 1) % weeks::weeks]
df = get_returns(weekly[['Adj Close', 'Dividend', 'Close']], 'simple')
df['Dividend'] = (weekly.Dividend.diff() / weekly.Close.shift(1))
df = df * 100 * 52 / weeks
from datetime import timedelta
ds = df.index
xlim = [
ds[0] - timedelta(days=3 * weeks), ds[-1] + timedelta(days=3 * weeks)
]
plt.figure(figsize=(14, 3))
plt.title("Annualized Return")
plt.hlines(xmin=xlim[0], xmax=xlim[1], y=0)
plt.hlines(xmin=xlim[0],
xmax=xlim[1],
y=df['Adj Close'].mean(),
linestyle='--',
color='#1f77b4')
plt.hlines(xmin=xlim[0],
xmax=xlim[1],
y=df.Dividend.mean(),
linestyle='--',
color='#ff7f0e')
plt.bar(ds, df.Close, width=5 * weeks, label='Yield')
plt.bar(ds,
df.Dividend,
bottom=df.Close,
width=5 * weeks,
label='Div_Yield')
plt.plot(ds, df['Adj Close'], 'o-', label='Adj_Yield')
plt.xlabel("Date to sell")
plt.xlim(xlim)
plt.ylim([np.min(df.values) - 0.2, np.max(df.values) + 0.2])
plt.legend(bbox_to_anchor=(1.01, 0.9), loc='upper left')
plt.grid()
df.index = df.index.date
print(np.round(df, 2))
plt.show()
return np.round(df, 2).T
def show_trend(ticker, start=None, end=None):
"""
Plot price change, return and volatility
input: ticker ticker of the stock/ETF
start start date, default '2000-01-01'
end end date, default current
Return correlation between return and volatility
"""
from invest.useful import convert_time
from invest.get_data import read_ETF
from invest.calculation import get_returns
from time_series.functions import moving_agg, resample
from basic.mathe import covariance, correlation
start, end = convert_time(start, end)
data = read_ETF(ticker)[start:end]
fig = plt.figure(figsize=(14,6))
fig.add_axes([0.05,0.68,0.94,0.3])
for c in ['Close','Adj Close']:
plt.plot(data.index, data[c], label=c)
plt.xlim(data.index[0],data.index[-1])
plt.xticks([])
plt.ylabel("Price ($)")
plt.legend(loc='best')
weekly = resample(data, style='week', method='close')
df = get_returns(weekly.Close, 'simple')
fig.add_axes([0.05,0.38,0.94,0.3])
m = moving_agg(df, window=52, step=1, func=np.sum)
plt.plot(df.index[51:], m*100)
plt.hlines(xmin=data.index[0], xmax=data.index[-1], y=0, linestyle='--')
plt.xlim(data.index[0],data.index[-1])
plt.xticks([])
plt.ylabel("Annual Return (%)")
plt.legend(loc='best')
fig.add_axes([0.05,0.08,0.94,0.3])
v = moving_agg(df, window=52, step=1, func=covariance)
v = np.sqrt(v*52)
plt.plot(df.index[51:], v*100)
plt.xlim(data.index[0],data.index[-1])
plt.ylabel("Volatility (%)")
plt.gca().set_ylim(bottom=0)
plt.legend(loc='best')
corr = correlation(m, v)
print("Correlation between return and volatility:", corr)
plt.show()
return corr
def generate(tickers, features):
"""
input: tickers a list of tickers of stocks/ETFs
features a list of feature names to be calculated
Return a pandas DataFrame
"""
from invest.get_data import read_ETF
from time_series.functions import resample
from invest.calculation import get_returns
data = pd.DataFrame()
for t in tickers:
etf = read_ETF(t)
weekly = resample(etf, style='week', method='close')
ret = get_returns(weekly, style='simple')
rv = get_return_vol(ret, scale=52, ret=True, plotit=False)
def set_benchmark(self):
ticker = self.benchmark.get().upper()
if ticker == "":
self._logger_.info("Remove benchmark")
self._data_['benchmark'] = 0
else:
self._logger_.info("Set benchmark as {}".format(ticker))
try:
data = read_ETF(ticker, file_dir=path + "\\data_temp")
except:
self._logger_.error("Cannot load benchmark {}".format(ticker))
else:
self._data_['benchmark'] = get_returns(resample(
data.Close, column=None, style="week", method='close'),
style='simple',
fillna=False)
self.update_plot()
def main_frame(self):
holdings = pd.read_csv(file_dir + '\\data\\holdings.csv',
index_col=None,
parse_dates=[0]).dropna()
data = {}
for t in holdings.Ticker.unique():
data[t] = read_ETF(t)
if (data[t].index[-1].date() !=
date.today()) and (date.today().weekday() < 5):
self._logger_.debug("Latest date is {}. Update local data for ticker {}"\
.format(data[t].index[-1].date(), t))
update_ETF_data([t],
dt="2000-01-01",
file_dir=file_dir + '\\data\\')
add_dividend(data[t],
price='Close',
adj='Adj Close',
out='Dividend')
def pop_up():
self._logger_.info("Popup window")
frame = tk.Tk()
self.transactions(frame)
frame.mainloop()
tk.Button(self, text='Add Transactions',
command=pop_up).pack(anchor=tk.E)
fig = plot_embed(self)
fig.pack()
table = tk.Frame(self)
table.pack()
for i, text in enumerate([
"Ticker", "Buy Date", "Buy Price", "Current Price",
"Buy Shares", "Reinvested Shares", "Capital Gain",
"Dividend Gain", "Total Gain"
]):
tk.Label(table, text=text).grid(row=0, column=i, padx=5, pady=5)
total_gain = 0
for loc, i in enumerate(holdings.index):
t = holdings.Ticker[i]
d = holdings.Date[i]
p = holdings.Price[i]
s = holdings.Shares[i]
r = (date.today() - d.date()).days / 365
print(r)
vals = [
t, "{}".format(d.date()), "{:.2f}".format(p),
"{:.2f}".format(data[t].iloc[-1, :].Close), "{:.0f}".format(s)
]
current_share = data[t].loc[d,
'Close'] / data[t].loc[d,
'Adj Close'] * s
vals.append("{:.4f}".format(current_share - s))
vals.append("{:.2f} ({:.2%})".format(
(data[t].iloc[-1, :].Close - p) * s,
(data[t].iloc[-1, :].Close / p - 1) / r))
vals.append("{:.2f} ({:.2%})".format(
data[t].loc[d:, 'Dividend'].sum() * s,
data[t].loc[d:, 'Dividend'].sum() / p / r))
gain = data[t].iloc[-1, :].Close * current_share - p * s
total_gain += gain
vals.append("{:.2f} ({:.2%})".format(gain, gain / p / s / r))
for j, text in enumerate(vals):
tk.Label(table, text=text).grid(row=loc + 1,
column=j,
padx=5,
pady=5)
tk.Label(table, text="Total gain is: {:.2f}".format(total_gain))\
.grid(row=holdings.shape[0]+1, column=0, columnspan=3)
def show_performance(ticker, current=None):
"""
Calculate total return and effective annual return for different starting point.
Simple return is calculated
input: ticker ticker of the stock/ETF
current the date to be calculated
Return a DataFrame with two rows: Total Return, Annual Return
"""
from datetime import date, timedelta
from invest.get_data import read_ETF
from time_series.functions import change_month
data = read_ETF(ticker)[['Close', 'Adj Close']]
if current is None:
current = data.index[-1]
else:
current = data[:current].index[-1]
mns = np.array([1, 3, 6, 12, 24, 36, 48, 60, 72, 84, 96, 108, 120])
dates = [change_month(current, -x) for x in mns]
df = pd.DataFrame(columns=['Close', 'Adj Close'])
dates = np.array(dates)
dates = dates[np.where(dates >= data.index[0].date())]
for d in dates:
vals = data[:d]
if vals.empty:
break
df.loc[vals.index[-1].date(), :] = 100 * (
data.loc[current, :] / vals.iloc[-1, :] - 1)
df['label'] = [
'1M', '3M', '6M', '1Y', '2Y', '3Y', '4Y', '5Y', '6Y', '7Y', '8Y', '9Y',
'10Y'
][:df.shape[0]]
df['years'] = mns[:df.shape[0]] / 12
result = pd.DataFrame(index=df.index)
result['Total Return'] = df.Close
result['Annual Return'] = df.Close / df.years
result['Adjusted Total Return'] = df['Adj Close']
result['Adjusted Annual Return'] = df['Adj Close'] / df.years
plt.figure(figsize=(14, 8))
plt.subplot(2, 1, 1)
plt.title("Performance for close price (not adjusted)")
plt.bar(df.index.map(lambda x: x - timedelta(days=12)),
df.Close,
width=24,
label='Total Return')
plt.bar(df.index.map(lambda x: x + timedelta(days=12)),
df.Close / df.years,
width=24,
label='Annual Return')
plt.hlines(xmin=df.index[-1] - timedelta(days=24), xmax=current, y=0)
plt.xticks(df.index, df.label)
plt.xlim(df.index[-1] - timedelta(days=24), current)
plt.ylim(max(-10,
np.min(result.values) - 1), min(20,
np.max(result.values) + 1))
plt.legend(loc='best')
plt.grid()
plt.subplot(2, 1, 2)
plt.title("Performance for close price (adjusted)")
plt.bar(df.index.map(lambda x: x - timedelta(days=12)),
df['Adj Close'],
width=24,
label='Total Return')
plt.bar(df.index.map(lambda x: x + timedelta(days=12)),
df['Adj Close'] / df.years,
width=24,
label='Annual Return')
plt.hlines(xmin=df.index[-1] - timedelta(days=24), xmax=current, y=0)
plt.xticks(df.index, df.label)
plt.xlim(df.index[-1] - timedelta(days=24), current)
plt.ylim(max(-10,
np.min(result.values) - 1), min(20,
np.max(result.values) + 1))
plt.legend(loc='best')
plt.grid()
print(np.round(result[::-1].astype(float), 2))
plt.show()
return np.round(result[::-1].astype(float), 2).T
