def optimize_portfolio(sd=dt.datetime(2008,1,1), ed=dt.datetime(2009,1,1), \
syms=['GOOG','AAPL','GLD','XOM'], gen_plot=False):
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
prices_all = prices_all.fillna(method='backfill')
prices_all = prices_all.fillna(method='ffill')
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
rfr = 0.0
sf = 252.0
# find the allocations for the optimal portfolio
# note that the values here ARE NOT meant to be correct for a test case
arg_num = len(prices.columns)
allocs = np.asarray(
arg_num *
[1 / float(arg_num)]) # add code here to find the allocations
#cr, adr, sddr, sr = [0.25, 0.001, 0.0005, 2.1] # add code here to compute stats
#arg_num=len(prices.columns)
def con(x):
return float(sum(x)) - 1.0
cons = ({'type': 'eq', 'fun': con})
bnds = arg_num * [(0.0, 1.0)]
min_result = spo.minimize(lambda x: prices.div(prices.iloc[0]).multiply(x).
sum(axis=1).pct_change().std(),
allocs,
method='SLSQP',
bounds=bnds,
constraints=cons)
allocs = min_result.x
prices_test = prices.div(prices.iloc[0])
prices_test1 = prices_test.multiply(allocs)
port_val = prices_test1.sum(axis=1)
cr = port_val.iloc[-1] / port_val.iloc[0] - 1
daily_ret = port_val.pct_change()
adr = daily_ret.mean()
sddr = daily_ret.std()
daily_rfr = rfr
sharp_ratio = (daily_ret.sub(daily_rfr)).mean() / (
(daily_ret.sub(daily_rfr)).std())
sr = sharp_ratio * np.power(sf, 1.0 / 2.0)
# Get daily portfolio value
#port_val = prices_SPY # add code here to compute daily portfolio values
# Compare daily portfolio value with SPY using a normalized plot
if gen_plot:
# add code to plot here
#df_temp = pd.concat([port_val, prices_SPY], keys=['Portfolio', 'SPY'], axis=1)
prices_SPY_test = prices_SPY.div(prices_SPY.iloc[0])
plot_prices = pd.DataFrame(
prices_test.multiply(allocs).sum(axis=1)).join(
pd.DataFrame(prices_SPY_test)).rename(columns={0: 'PORTFOLIO'})
plot_data(plot_prices)
pass
return allocs, cr, adr, sddr, sr
def BestPossible():
# Read data
dates = pd.date_range('2008-1-1', '2009-12-31')
start_val = 100000
syms = ['AAPL']
prices_all = get_data(syms, dates)
prices_port = prices_all[syms] # only price of each stock in portfolio
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
dfprices = pd.DataFrame(prices_port)
dfprices['Cash'] = start_val - 200 * dfprices.ix[0, 'AAPL']
dfprices['Holding'] = 200
dfprices['Stock values'] = dfprices['Holding'] * dfprices['AAPL']
dfprices['Port_val'] = dfprices['Cash'] + dfprices['Stock values']
dfprices['Benchmark'] = dfprices['Port_val'] / dfprices['Port_val'].ix[
0, :]
dfprices['Best'] = Best_strategy(prices_all[syms])
dfprices['Best possible portfolio'] = dfprices['Best'] / dfprices[
'Best'].ix[0, :]
df = pd.concat(
[dfprices['Benchmark'], dfprices['Best possible portfolio']],
keys=['Benchmark', 'Best possible portfolio'],
axis=1)
plot_data(df,
title="Benchmark vs. Best possible portfolio",
ylabel="Normalized price",
xlabel="Date")
crBM, adrBM, sddrBM = compute_portfolio_stats(dfprices['Benchmark'])
crBest, adrBest, sddrBest = compute_portfolio_stats(
dfprices['Best possible portfolio'])
print "statistics of Benchmark", crBM, adrBM, sddrBM
print "statistics of Best Stategy", crBest, adrBest, sddrBest
return df
def get_RSI(prices, gen_plot=False):
prices_RSI = prices.copy()
prices_RSI = prices_RSI / prices_RSI.iloc[0]
prices_RSI['gain/loss'] = prices_RSI['JPM'] - prices_RSI['JPM'].shift(1)
prices_RSI['gain'] = prices_RSI['gain/loss'].apply(
lambda x: x > 0 and x or 0)
prices_RSI['loss'] = prices_RSI['gain/loss'].apply(
lambda x: x < 0 and x or 0)
list_gain = prices_RSI['gain'].tolist()
list_loss = prices_RSI['loss'].tolist()
t = 7
list_RS = []
for i in range(0, t):
list_RS.append(np.NAN)
for i in range(t, prices.shape[0]):
RS = (np.mean(list_gain[0:i]) *
(t - 1) + list_gain[i]) / (np.mean(list_loss[0:i]) *
(t - 1) + list_loss[i]) * (-1)
list_RS.append(RS)
prices_RSI['RS'] = list_RS
prices_RSI['RSI'] = prices_RSI['RS'].apply(lambda x: 100 - 100 / (1 + x))
#plot graph
if gen_plot:
print("Technical Analysis by RSI: ")
plot_data(prices_RSI[['RSI']],
title="Method-4 RSI",
xlabel="Date",
ylabel="Price")
plt.savefig('indicator_RSI')
"""
def assess_portfolio(sd=dt.datetime(2008, 1, 1), ed=dt.datetime(2009, 1, 1),
syms=['GOOG', 'AAPL', 'GLD', 'XOM'],
allocs=[0.1, 0.2, 0.3, 0.4],
sv=1000000, rfr=0.0, sf=252.0,
gen_plot=False):
if sum(allocs) != 1:
raise Exception('Allocations to the equities should sum to 1')
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates)
prices = prices_all[syms]
prices_SPY = prices_all['SPY']
normed = prices / prices.iloc[0]
alloced = normed * allocs
pos_vals = alloced * sv
port_val = pos_vals.sum(axis=1)
cr, adr, sddr, sr = compute_portfolio_stats(port_val=port_val, rfr=rfr, sf=sf)
if gen_plot:
df_temp = pd.concat([port_val, prices_SPY], keys=['Portfolio', 'SPY'], axis=1)
df_temp = df_temp / df_temp.iloc[0]
plot_data(df_temp)
ev = port_val[-1]
return cr, adr, sddr, sr, ev
def optimize_portfolio(sd=dt.datetime(2008,1,1), ed=dt.datetime(2009,1,1), \
syms=['GOOG','AAPL','GLD','XOM'], gen_plot=False):
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
# find the allocations for the optimal portfolio
# note that the values here ARE NOT meant to be correct for a test case
def r_sharpratio(allocs):
return analysis(prices, allocs)[3] * (-1)
xGuess = np.ones(len(syms)) / float(len(syms))
bs = [(0, 1) for i in range(len(syms))]
cs = ({ 'type': 'eq', 'fun': lambda x: 1.0 - np.sum(x) })
allocs = spo.minimize(r_sharpratio, xGuess, method = 'SLSQP', bounds = bs, constraints = cs)
cr, adr, sddr, sr = analysis(prices, allocs.x)
# Get daily portfolio value
port_val = prices_SPY / prices_SPY.ix[0] # add code here to compute daily portfolio values
port_val2 = (prices / prices.ix[0] * allocs.x).sum(axis = 1)
# Compare daily portfolio value with SPY using a normalized plot
if gen_plot:
# add code to plot here
df = pd.concat([port_val2, port_val], keys=['Portfolio', 'SPY'], axis=1)
plot_data(df, title="Daily Portfolio Value and SPY", xlabel="Date", ylabel="Price")
return allocs.x, cr, adr, sddr, sr
def test_code1():
# this is a helper function you can use to test your code
# note that during autograding his function will not be called.
# Define input parameters
of = "/home/mdonaher/PycharmProjects/mc2_pc2/orders.csv"
sv = 10000
start_date = dt.datetime(2007, 12, 31)
end_date = dt.datetime(2009, 12, 31)
# Process orders
portvals = compute_portvals(
orders_file=of,
start_val=sv,
sd=start_date,
ed=end_date,
)
print type(portvals)
if isinstance(portvals, pd.DataFrame):
portvals = portvals[portvals.columns[0]] # just get the first column
else:
print "warning, Your code did not return a DataFrame"
plot_data(portvals)
# Get portfolio stats
# Here we just fake the data. you should use your code from previous assignments.
start_date = portvals.index.min()
end_date = portvals.index.max()
start_date1 = ((portvals.index.min()).to_pydatetime()).strftime('%Y-%m-%d')
end_date1 = ((portvals.index.max()).to_pydatetime()).strftime('%Y-%m-%d')
calc_cum_ret = (portvals.ix[end_date, ['Cummulative_Value']] /
portvals.ix[start_date, ['Cummulative_Value']]) - 1
calc_adr = ((portvals / portvals.shift()) - 1).mean()
calc_sddr = ((portvals / portvals.shift()) - 1).std()
cum_ret, avg_daily_ret, std_daily_ret, sharpe_ratio = [
calc_cum_ret, calc_adr, calc_sddr, 1.5
]
cum_ret_SPY, avg_daily_ret_SPY, std_daily_ret_SPY, sharpe_ratio_SPY = [
0.2, 0.01, 0.02, 1.5
]
# Compare portfolio against $SPX
print "Date Range: {} to {}".format(start_date1, end_date1)
print
print "Sharpe Ratio of Fund: {}".format(sharpe_ratio)
print "Sharpe Ratio of SPY : {}".format(sharpe_ratio_SPY)
print
print "Cumulative Return of Fund: {}".format(cum_ret)
print "Cumulative Return of SPY : {}".format(cum_ret_SPY)
print
print "Standard Deviation of Fund: {}".format(std_daily_ret)
print "Standard Deviation of SPY : {}".format(std_daily_ret_SPY)
print
print "Average Daily Return of Fund: {}".format(avg_daily_ret)
print "Average Daily Return of SPY : {}".format(avg_daily_ret_SPY)
print
print "Final Portfolio Value: {}".format(portvals[-1])
def optimize_portfolio(sd, ed, syms, gen_plot=True):
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
prices_all.fillna(method="ffill", inplace=True)
prices_all.fillna(method="bfill", inplace=True)
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
allocs = find_optimum_allocs(prices)
allocs = allocs / np.sum(allocs)
port_val = get_portfolio_value(prices, allocs)
cum_ret, avg_daily_ret, std_daily_ret, sharp_ratio = get_portfolio_status(
port_val)
normed_SPY = prices_SPY / prices_SPY.ix[0, :]
if gen_plot:
df_temp = pd.concat([port_val, normed_SPY],
keys=['Portfolio', 'SPY'],
axis=1)
plot_data(df_temp, title='Daily Portfolio Value and SPY')
pass
return allocs, cum_ret, avg_daily_ret, std_daily_ret, sharp_ratio
def main():
dates = pd.date_range('2009-01-01', '2012-12-31')
symbols = ['SPY']
df = get_data(symbols, dates)
plot_data(df)
daily_returns = compute_daily_returns(df)
# histogram
daily_returns.hist(bins=20)
'''
call this twice if wanting to plot 2+ charts on same chart:
daily_returns['SPY'].hist(bins=20, label="SPY")
daily_returns['XOM'].hist(bins=20, label="XOM")
'''
mean = daily_returns['SPY'].mean()
std = daily_returns['SPY'].std()
plt.axvline(mean, color='w', linestyle='dashed', linewidth=2)
plt.axvline(std, color='r', linestyle='dashed', linewidth=2)
plt.axvline(-std, color='r', linestyle='dashed', linewidth=2)
plt.show()
print daily_returns.kurtosis()
def test_run():
start_date = '2009-01-01'
end_date = '2012-12-31'
dates = pd.date_range(start_date, end_date)
symbols = ['SPY']
df = get_data(symbols, dates)
daily_returns = compute_daily_returns(df)
plot_data(daily_returns, title="Daily returns", ylabel="Daily returns")
daily_returns.hist(bins=20)
mean = daily_returns['SPY'].mean()
print("mean:\t\t\t{}".format(mean))
std = daily_returns['SPY'].std()
print("std. deviation:\t{}".format(std))
plt.axvline(mean, color='red', linestyle='dashed', linewidth=1)
plt.axvline(std, color='black', linestyle='dashed', linewidth=1)
plt.axvline(-std, color='black', linestyle='dashed', linewidth=1)
plt.show()
kurtosis = daily_returns.kurtosis()
print("kurtosis:\t{}".format(kurtosis))
def assess_portfolio(start_date, end_date, symbols, allocs, start_val=1):
"""Simulate and assess the performance of a stock portfolio."""
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(start_date, end_date)
prices_all = get_data(symbols, dates) # automatically adds SPY
prices = prices_all[symbols] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
# Get daily portfolio value
port_val = get_portfolio_value(prices, allocs, start_val)
plot_data(port_val, title="Daily Portfolio Value")
# Get portfolio statistics (note: std_daily_ret = volatility)
cum_ret, avg_daily_ret, std_daily_ret, sharpe_ratio = get_portfolio_stats(port_val)
# Print statistics
print "Start Date:", start_date
print "End Date:", end_date
print "Symbols:", symbols
print "Allocations:", allocs
print "Sharpe Ratio:", sharpe_ratio
print "Volatility (stdev of daily returns):", std_daily_ret
print "Average Daily Return:", avg_daily_ret
print "Cumulative Return:", cum_ret
# Compare daily portfolio value with SPY using a normalized plot
df_temp = pd.concat([port_val, prices_SPY], keys=['Portfolio', 'SPY'], axis=1)
plot_normalized_data(df_temp, title="Daily portfolio value and SPY")
def optimize_portfolio(sd=dt.datetime(2008, 1, 1), ed=dt.datetime(2009, 1, 1),
syms=['GOOG', 'AAPL', 'GLD', 'XOM'], gen_plot=False):
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
# find the allocations for the optimal portfolio
allocs = optimize_sharpe_ratio(prices)
# compute stats
cr, adr, sddr, sr = compute_portfolio_stats(prices, allocs)
# Get daily portfolio value
normalized_prices = prices / prices.ix[0]
alloced = normalized_prices * allocs
port_val = alloced.sum(axis=1)
# Compare daily portfolio value with SPY using a normalized plot
if gen_plot:
# add code to plot here
df_temp = pd.concat([port_val, prices_SPY],
keys=['Portfolio', 'SPY'], axis=1)
df_temp = df_temp / df_temp.ix[0]
plot_data(df_temp, "Daily portfolio value and SPY",
'Normalized price', 'Time')
return allocs, cr, adr, sddr, sr
def assess_portfolio(sd = dt.datetime(2008,1,1), ed = dt.datetime(2009,1,1), \
syms = ['GOOG','AAPL','GLD','XOM'], \
allocs=[0.1,0.2,0.3,0.4], \
sv=1000000, rfr=0.0, sf=252.0, \
gen_plot=False):
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
# compute daily portfolio values
port_val = compute_portfolio_value(prices, allocs, start_val=1000000)
# compute portfolio statistics (note: std_daily_ret = volatility)
cr, adr, sddr, sr = compute_portfolio_statistics(port_val, freq_sample=252)
# Compare daily portfolio value with SPY using a normalized plot
if gen_plot:
# add code to plot here
df_temp = pd.concat([port_val, prices_SPY],
keys=['Portfolio', 'SPY'],
axis=1)
plot_data(df_temp)
pass
return cr, adr, sddr, sr
def test_code():
# see format of [order].csv in orders directory
of = "./orders/orders2.csv"
sv = 1000000
# Process orders
portvals = compute_portvals(orders_file=of, start_val=sv)
print portvals
if isinstance(portvals, pd.DataFrame):
portvals = portvals[portvals.columns[0]] # just get the first column
else:
"warning, code did not return a DataFrame"
# plot with spy to compare
start_date = portvals.index[0]
end_date = portvals.index[-1]
dates = pd.date_range(start_date, end_date)
temp = get_data(['GOOG'], dates)
SPY = temp['SPY']
normalized_prices_SPY = SPY.copy()
normalized_prices_SPY = (normalized_prices_SPY) / SPY[0]
normalized_portval = portvals.copy()
normalized_portval = (normalized_portval) / portvals[0]
df_temp = pd.concat([normalized_portval, normalized_prices_SPY],
keys=['Portfolio', 'SPY'],
axis=1)
plot_data(df_temp)
def assess_portfolio(sd = dt.datetime(2008,1,1), ed = dt.datetime(2009,1,1), \
syms = ['GOOG','AAPL','GLD','XOM'], \
allocs=[0.1,0.2,0.3,0.4], \
sv=1000000, rfr=0.0, sf=252.0, \
gen_plot=False):
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
prices_SPY = (prices_SPY / prices_SPY.iloc[0]) * sv
# Get daily portfolio value
port_val = get_portfolio_value(prices, allocs, sv)
# Get portfolio statistics (note: std_daily_ret = volatility)
cr, adr, sddr, sr = get_portfolio_stats(port_val, rfr, sf)
# Compare daily portfolio value with SPY using a normalized plot
if gen_plot:
# Plot normalized portfolio value.
df_temp = pd.concat([port_val / sv, prices_SPY / sv],
keys=['Portfolio', 'SPY'],
axis=1)
plot_data(df_temp,
title="Daily portfolio value and SPY",
ylabel="Normalized price")
# Compute end value
ev = port_val[-1]
return cr, adr, sddr, sr, ev
def optimize_portfolio(start_date, end_date, symbols):
"""Simulate and optimize portfolio allocations."""
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(start_date, end_date)
prices_all = get_data(symbols, dates) # automatically adds SPY
prices = prices_all[symbols] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
# Get optimal allocations
allocs = find_optimal_allocations(prices)
allocs = allocs / np.sum(allocs) # normalize allocations, if they don't sum to 1.0
# Get daily portfolio value (already normalized since we use default start_val=1.0)
port_val = get_portfolio_value(prices, allocs)
# Get portfolio statistics (note: std_daily_ret = volatility)
cum_ret, avg_daily_ret, std_daily_ret, sharpe_ratio = get_portfolio_stats(port_val)
# Print statistics
print "Start Date:", start_date
print "End Date:", end_date
print "Symbols:", symbols
print "Optimal allocations:", allocs
print "Sharpe Ratio:", sharpe_ratio
print "Volatility (stdev of daily returns):", std_daily_ret
print "Average Daily Return:", avg_daily_ret
print "Cumulative Return:", cum_ret
# Compare daily portfolio value with normalized SPY
normed_SPY = prices_SPY / prices_SPY.ix[0, :]
df_temp = pd.concat([port_val, normed_SPY], keys=['Portfolio', 'SPY'], axis=1)
plot_data(df_temp, title="Daily Portfolio Value and SPY")
def assess_portfolio(sd = dt.datetime(2008,1,1), ed = dt.datetime(2009,1,1), \
syms = ['GOOG','AAPL','GLD','XOM'], \
allocs=[0.1,0.2,0.3,0.4], \
sv=1000000, rfr=0.0, sf=252.0, \
gen_plot=False):
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
# Get daily portfolio value
# Get portfolio statistics (note: std_daily_ret = volatility)
cr, adr, sddr, sr = [0.25, 0.001, 0.0005,
2.1] # add code here to compute stats
port_val = prices_SPY # add code here to compute daily portfolio values
normed = prices / prices.iloc[0]
#print (allocs)
alloced = normed * allocs
pos_vals = alloced * sv
port_val = pos_vals.sum(axis=1)
normed_SPY = prices_SPY / prices_SPY.iloc[0]
daily_rets = port_val / port_val.shift(
1) - 1 #by 1 row, subtracted by 1 to view change
ev = sv
ev = port_val[len(pos_vals.index) - 1]
cr = ev / port_val[0] - 1 #portval.length?
adr = daily_rets.mean() #check that avg. daily return is correct
sddr = daily_rets.std()
sr_annualized = math.sqrt(sf)
sr = sr_annualized * (adr - rfr) / (sddr - rfr)
# Compare daily portfolio value with SPY using a normalized plot
if gen_plot:
# add code to plot here
#plot_data( normed, title="Normalized Stock Prices", ylabel="Normalized price")
#df_temp = pd.concat([port_val, prices_SPY], keys=['Portfolio', 'SPY'], axis=1)
df_temp = pd.concat([port_val / sv, normed_SPY],
keys=['Portfolio', 'SPY'],
axis=1)
#df_temp = pd.concat([normed_SPY], keys=['Portfolio', 'SPY'], axis=1)
#df_temp = pd.concat([port_val], keys=['Portfolio', 'SPY'], axis=1)
#plot_data([port_val, prices_SPY], title="Daily portfolio value and SPY", ylabel="Normalized price")
plot_data(df_temp,
title='Daily portfolio value and SPY',
ylabel='Normalized price')
pass
# Add code here to properly compute end value
return cr, adr, sddr, sr, ev
def test_run():
# Read data
dates = pd.date_range('2009-01-01', '2012-12-31')
symbols = ['SPY']
df = get_data(symbols, dates)
plot_data(df)
# Compute daily returns
daily_returns = compute_daily_returns(df)
plot_data(daily_returns, title="Daily returns", ylabel="Daily returns")
# Plot a histogram
daily_returns.hist(bins=20) # changing # of bins to 20
# Get mean and standard deviation
mean = daily_returns['SPY'].mean()
print "mean =", mean
std = daily_returns['SPY'].std()
print "std =", std
plt.axvline(mean,color='w',linestyle='dashed',linewidth=2)
plt.axvline(std,color='r',linestyle='dashed',linewidth=2)
plt.axvline(-std,color='r',linestyle='dashed',linewidth=2)
plt.show()
# Compute kurtosis
print daily_returns.kurtosis()
def SMA(self, sd, ed, stock, SMA_Nday):
dates = pd.date_range(sd, ed)
prices1 = ut.get_data([stock], dates)
nsd = sd - dt.timedelta(days=100)
ndates = pd.date_range(nsd, ed)
prices = ut.get_data([stock], ndates)
prices_apple = prices[[stock]]
# plot_data(prices)
SMA_apple = pd.rolling_mean(prices_apple,
SMA_Nday,
min_periods=1,
center=False)
price_SMA = prices_apple / SMA_apple
price_SMA_norm = (price_SMA - price_SMA.mean()) / price_SMA.std()
# price_SMA = price_SMA/price_SMA.ix[SMA_Nday]
if False:
SMA_apple = SMA_apple / prices_apple.ix[0]
prices_n = prices_apple / prices_apple.ix[0]
SMA_all = pd.concat([prices_n, SMA_apple, price_SMA],
keys=['Price', 'SMA', 'Price/SMA'],
axis=1)
ut.plot_data(SMA_all,
title='Technical indicator: SMA',
ylabel='Relative Value')
return pd.DataFrame(price_SMA_norm, index=prices1.index).fillna(0)
def test_run():
# Read data
dates = pd.date_range('2009-01-01', '2012-12-31')
symbols = ['SPY']
df = get_data(symbols, dates)
plot_data(df)
# Compute daily returns
daily_returns = compute_daily_returns(df)
plot_data(daily_returns, title="Daily returns", ylabel="Daily returns")
# Plot a histogram
daily_returns.hist(bins=20) # changing # of bins to 20
# Get mean and standard deviation
mean = daily_returns['SPY'].mean()
print "mean =", mean
std = daily_returns['SPY'].std()
print "std =", std
plt.axvline(mean, color='w', linestyle='dashed', linewidth=2)
plt.axvline(std, color='r', linestyle='dashed', linewidth=2)
plt.axvline(-std, color='r', linestyle='dashed', linewidth=2)
plt.show()
# Compute kurtosis
print daily_returns.kurtosis()
def assess_portfolio(sd = dt.datetime(2008,1,1), ed = dt.datetime(2009,1,1), \
syms = ['GOOG','AAPL','GLD','XOM'], \
allocs=[0.1,0.2,0.3,0.4], \
sv=1000000, rfr=0.0, sf=252.0, \
gen_plot=False):
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
prices_SPY = (prices_SPY/prices_SPY.iloc[0])*sv
# Get daily portfolio value
port_val = get_portfolio_value(prices, allocs, sv)
# Get portfolio statistics (note: std_daily_ret = volatility)
cr, adr, sddr, sr = get_portfolio_stats(port_val, rfr, sf)
# Compare daily portfolio value with SPY using a normalized plot
if gen_plot:
# Plot normalized portfolio value.
df_temp = pd.concat([port_val/sv, prices_SPY/sv], keys=['Portfolio', 'SPY'], axis=1)
plot_data(df_temp,
title="Daily portfolio value and SPY",
ylabel="Normalized price")
# Compute end value
ev = port_val[-1]
return cr, adr, sddr, sr, ev
def assess_portfolio(portfolio=None,sd = dt.datetime(2008,1,1), ed = dt.datetime(2009,1,1), \
syms = ['GOOG','AAPL','GLD','XOM'], \
allocs=[0.1,0.2,0.3,0.4], \
sv=1000000, rfr=0.0, sf=252.0,precios=0, \
gen_plot=False):
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(sd, ed)
if(type(portfolio)==type(None)):
prices_all = get_data(syms, dates) # automatically adds SPY
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
else:
#pdb.set_trace()
prices_all = get_data(syms,dates)
prices = portfolio# pd.concat([portfolio,prices_all[syms]],axis=1)
prices_SPY = prices_all[syms]
# Get daily portfolio value
if( (isinstance(precios,pd.core.series.Series))):
pdb.set_trace()
precios = precios.to_frame()
normalized_values = normalize_data(precios)
else:
normalized_values = normalize_data(prices)
alloced_values = normalized_values*allocs
pos_val = alloced_values*sv
port_val = pos_val.sum(axis=1) # add code here to compute daily portfolio values
d_returns = daily_returns(port_val)
cr = port_val.iloc[-1]/port_val.iloc[0] -1
adr = d_returns.mean()
sddr = d_returns.std()
sr = (adr/sddr)*np.sqrt(252)
prices_SPY = normalize_data(prices_SPY)
port_val = normalize_data(port_val)
port_val = pd.DataFrame(data=port_val,index=port_val.index,columns=['Portfolio'])
# Get portfolio statistics (note: std_daily_ret = volatility)
#cr, adr, sddr, sr = [0.25, 0.001, 0.0005, 2.1] # add code here to compute stats
# Compare daily portfolio value with SPY using a normalized plot
if gen_plot:
# add code to plot here
if(type(portfolio)==type(None)):
df_temp = pd.concat([port_val, prices_SPY], keys=['Portfolio', 'SPY'], axis=1)
else:
#pdb.set_trace()
df_temp = pd.concat([port_val, prices_SPY], axis=1)
plot_data(df_temp)
# Add code here to properly compute end value
ev = sv
return(cr, adr, sddr, sr, ev)
def RSI(self, sd, ed, stock, SMA_Nday):
dates = pd.date_range(sd, ed)
prices1 = ut.get_data([stock], dates)
nsd = sd - dt.timedelta(days=100)
ndates = pd.date_range(nsd, ed)
prices = ut.get_data([stock], ndates)
prices_apple = prices[[stock]]
price_delta = prices_apple.diff()
dUp = price_delta.copy()
dDown = price_delta.copy()
dUp[dUp < 0] = 0
dDown[dDown > 0] = 0
rUp = pd.rolling_mean(dDown, SMA_Nday, min_periods=1,
center=False).abs()
rDown = pd.rolling_mean(dUp, SMA_Nday, min_periods=1,
center=False).abs()
RS = rUp / rDown
RSI = 100 - 100 / (1 + RS)
RSI_norm = (RSI - RSI.mean()) / RSI.std()
if False:
prices_1 = prices_apple / prices_apple.ix[0]
RSI_1 = RSI / RSI.ix[SMA_Nday]
RSI_all = pd.concat([prices_1, RSI_1],
keys=['Price', 'RSI'],
axis=1)
plt.figure(3)
ut.plot_data(RSI,
title='Technical indicator: Relative Strength Index',
ylabel='RSI')
return pd.DataFrame(RSI_norm, index=prices1.index).fillna(0)
def optimize_portfolio(sd=dt.datetime(2008,1,1), ed=dt.datetime(2009,1,1), \
syms=['GOOG','AAPL','GLD','XOM'], gen_plot=False):
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
# find the allocations for the optimal portfolio
# note that the values here ARE NOT meant to be correct for a test case
# add code here to find the allocations
x0 = np.random.random(len(syms))
x0 /= x0.sum()
# x0=np.asarray([0.2, 0.2, 0.3, 0.3, 0.0])
fun = lambda x: -sharp_ratio(prices.values,x)
cons = ({ 'type': 'eq', 'fun': lambda inputs: 1 - np.sum(inputs) })
bnds = tuple((0,None) for i in range(len(syms)))
res = minimize(fun, x0 , method='SLSQP', bounds=bnds, constraints=cons)
allocs = res.x
cr, adr, sddr, sr = [0.25, 0.001, 0.0005, 2.1] # add code here to compute stats
priceSPY=prices_SPY.values
priceSPY /= priceSPY[0]
price_stocks = prices.values
price_stocks /= price_stocks[0]
price_stocks *= allocs
port_val = pd.DataFrame(price_stocks.sum(axis=1),index=prices.index)
prices_SPY = pd.DataFrame(priceSPY,index=prices.index)
# Get daily portfolio value
# port_val = prices_SPY # add code here to compute daily portfolio values
cr = port_val.values[-1] -1
dr = port_val.values
drShift = np.vstack([dr[0],dr[0:(len(dr)-1)]])
dr = dr/drShift -1
adr = dr.mean()
sddr=dr.std()
k = math.sqrt(252)
sr = k*np.mean(dr)/np.std(dr)
# Compare daily portfolio value with SPY using a normalized plot
if gen_plot:
# add code to plot here
df_temp = pd.concat([port_val, prices_SPY], keys=['Portfolio', 'SPY'], axis=1)
df_temp.columns=df_temp.columns.get_level_values(0)
plot_data(df_temp, "Daily portfolio value and SPY", "Date", "Normalized prices")
return allocs, cr, adr, sddr, sr
def assess_portfolio(sd = dt.datetime(2008,1,1), ed = dt.datetime(2009,1,1),
syms = ['GOOG','AAPL','GLD','XOM'],
allocs=[0.1,0.2,0.3,0.4],
sv=1000000, rfr=0.0, sf=252.0,
gen_plot=False):
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
# Get portfolio statistics (note: std_daily_ret = volatility)
port_val = get_portfolio_value(prices, allocs, sv)
cr, adr, sddr, sr = compute_portfolio_stats(port_val, rfr, sf) # add code here to compute stats
# Compare daily portfolio value with SPY using a normalized plot
if gen_plot:
# add code to plot here
port_val_norm = port_val / port_val.iloc[0]
prices_SPY_norm = prices_SPY / prices_SPY.iloc[0]
df_temp = pd.concat([port_val_norm, prices_SPY_norm], keys=['Portfolio', 'SPY'], axis=1)
plot_data(df_temp)
pass
# Add code here to properly compute end value
ev = port_val[-1]
return cr, adr, sddr, sr, ev
def optimize_portfolio(sd=dt.datetime(2008,1,1), ed=dt.datetime(2009,1,1), \
syms=['GOOG','AAPL','GLD','XOM'], gen_plot=False):
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
normed_SPY = prices_SPY/prices_SPY.ix[0]
# find the allocations for the optimal portfolio
# note that the values here ARE NOT meant to be correct for a test case
noa = len(syms)
noa = (noa * [1. / noa,])
cons = ({'type': 'eq', 'fun': lambda x: np.sum(x) - 1})
bnds = tuple((0,1) for x in range(len(syms)))
min = co.minimize(negative_sharpe, noa, args=(prices,), method='SLSQP', bounds=bnds, constraints=cons)
allocs = np.asarray(min.x)
cr, adr, sddr, sr, normed_daily_portfolio = assess_portfolio(sd,ed,syms, allocs.tolist())
# Compare daily portfolio value with SPY using a normalized plot
if gen_plot:
df_temp = pd.concat([normed_daily_portfolio, normed_SPY], keys=['Portfolio', 'SPY'], axis=1)
plot_data(df_temp, title="Daily Portfolio Value and SPY", ylabel="Price")
plt.show()
pass
return allocs, cr, adr, sddr, sr
def assess_portfolio(start_date, end_date, symbols, allocs, start_val=1):
"""Simulate and assess the performance of a stock portfolio."""
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(start_date, end_date)
prices_all = get_data(symbols, dates) # automatically adds SPY
prices = prices_all[symbols] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
# Get daily portfolio value
port_val = get_portfolio_value(prices, allocs, start_val)
plot_data(port_val, title="Daily Portfolio Value")
# Get portfolio statistics (note: std_daily_ret = volatility)
cum_ret, avg_daily_ret, std_daily_ret, sharpe_ratio = get_portfolio_stats(port_val)
# Print statistics
print "Start Date:", start_date
print "End Date:", end_date
print "Symbols:", symbols
print "Allocations:", allocs
print "Sharpe Ratio:", sharpe_ratio
print "Volatility (stdev of daily returns):", std_daily_ret
print "Average Daily Return:", avg_daily_ret
print "Cumulative Return:", cum_ret
# Compare daily portfolio value with SPY using a normalized plot
df_temp = pd.concat([port_val, prices_SPY], keys=['Portfolio', 'SPY'], axis=1)
plot_normalized_data(df_temp, title="Daily portfolio value and SPY")
def optimize_portfolio(sd = dt.datetime(2008, 1, 1), ed = dt.datetime(2009, 1, 1), \
syms=['GOOG', 'AAPL', 'GLD', 'XOM'], gen_plot = False):
startValue = 1000000
tradeFrequency = 252.0
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
# Fill in missing data
prices_all.fillna(method="ffill")
prices_all.fillna(method="bfill")
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
# Find the allocations for the optimal portfolio
# Initialize allocations array with same allocation for each stock (initial guess) & normalize prices to first entries
allocs = np.asarray([(1.0 / len(syms)) for i in range(len(syms))])
normalized = prices / prices.values[0]
# Optimize allocation to minimize negative of Sharpe ratio equation (maximize Sharpe ratio)
bounds = [(0.0, 1.0) for i in range(len(syms))]
constraints = ({'type': 'eq', 'fun': lambda inputs: 1.0 - np.sum(inputs)})
minimized = opt.minimize(negSharpe,
allocs,
args=(normalized, startValue),
method='SLSQP',
constraints=constraints,
bounds=bounds,
options={'disp': True})
# Calculate portfolio value to use in calculation of return values
new_allocs = minimized.x
new_sddr = minimized.fun
allocations = normalized.multiply(new_allocs)
position_vals = allocations.multiply(startValue)
portfolio_val = position_vals.sum(axis=1)
# Get daily portfolio values, cumulative value, and Sharpe ratio
daily = (portfolio_val / portfolio_val.shift(1)) - 1
cumulative = (portfolio_val[-1] / portfolio_val[0]) - 1
sharpe = np.sqrt(tradeFrequency) * daily.mean() / new_sddr
# Compare daily portfolio value with SPY using a normalized plot
if gen_plot:
portfolio_val /= portfolio_val[0]
prices_SPY /= prices_SPY[0]
df_temp = pd.concat([portfolio_val, prices_SPY],
keys=['Portfolio', 'SPY'],
axis=1)
plot_data(df_temp,
title="Optimal Portfolio vs. SPY Daily Comparison",
ylabel="Normalized Price")
plt.savefig('plot.png')
pass
return new_allocs, cumulative, daily.mean(), new_sddr, sharpe
def test_run():
# Get initial data
start_date = '2004-08-31'
end_date = '2016-01-01'
dates = pd.date_range(start_date, end_date)
symbols = ['SPY', 'VOO', 'GLD', 'XOM', 'AAPL']
original_allocation = [0.2, 0.2, 0.2, 0.2, 0.2]
start_val = 20000 # one million dollars!!!
# plot the original data
original_data = util.portfolio_daily_values(start_val, dates, symbols,
original_allocation)
util.plot_data(original_data, "Original", label='Original')
# Plot the new data
new_allocation = fit_line(original_data, negative_sharpe, start_val, dates,
symbols, original_allocation)
new_data = util.portfolio_daily_values(start_val, dates, symbols,
new_allocation)
util.plot_data(new_data, "Optimized", label='Optimized')
plt.legend(loc='upper right')
print("Original Allocation=", original_allocation)
print("New allocation=", new_allocation)
plt.show()
def test_run():
symbols = ['SPY', 'XOM', 'GOOG', 'GLD']
dates = pd.date_range('2010-01-01', '2010-12-31')
df = get_data(symbols, dates)
plot_data(df)
#plot_selected(df, ['SPY', 'GOOG', 'IBM', 'GLD'], '2010-03-01', '2010-04-01')
print(df.mean())
def test_run():
# Read data
dates = pd.date_range('2009-01-01', '2012-12-31')
symbols = ['SPY', 'XOM']
df = get_data(symbols, dates)
plot_data(df)
""" Two separate histograms ==========="""
# Compute daily returns
daily_returns = compute_daily_returns(df)
plot_data(daily_returns, title="Daily returns", ylabel="Daily returns")
# Plot a histogram
daily_returns.hist(bins=20)
plt.show()
""" Histograms on the same graph ======"""
# Compute daily returns
daily_returns = compute_daily_returns(df)
# Compute and plot both histograms on the same chart
daily_returns['SPY'].hist(bins=20, label="SPY")
daily_returns['XOM'].hist(bins=20, label="XOM")
plt.legend(loc='upper right')
plt.show()
def generateMACDChart(price, a, b, c):
#emas with price
macd_line, signal_line, macd_histogram, ema_upper, ema_lower = macd(
price, a, b, c)
macd_data = pd.merge(price,
ema_upper,
left_index=True,
right_index=True)
macd_data = pd.merge(macd_data,
ema_lower,
left_index=True,
right_index=True)
macd_data = macd_data.ix[20:]
macd_data = macd_data / macd_data.ix[0]
macd_data.columns = [
'Price Normalized', '26 Day EMA Normalized Price',
'12 Day EMA Normalized Price'
]
#macd indicators
macd_line = pd.merge(macd_line,
signal_line,
left_index=True,
right_index=True)
#macd_line=pd.merge(macd_line, macd_histogram, left_index=True,right_index=True)
macd_line.columns = ['Signal Line', 'MACD Line']
plot_data(macd_data,
title="EMA Prices Over Time",
xlabel="Date",
ylabel="Price")
plot_data(macd_line,
title="MACD and Signal Line Over Time",
xlabel="Date",
ylabel="MACD or Signal Value")
def test_run():
# Read Data
dates = pd.date_range('2009-01-01', '2012-12-31')
symbols = ['SPY']
df = get_data(symbols, dates)
plot_data(df)
# Compute Daily Returns
daily_return = compute_daily_returns(df)
plot_data(daily_returns)
# Plot a histogram
daily_returns.hist() # Default number of bins is 10
daily_returns.hist(bins=20) # 20 bins
# Computing Histogram Statistics
mean = daily_returns['SPY'].mean()
std = daily_returns['SPY'].std()
plt.axvline(mean)
plt.axvline(std)
plt.axvline(-std)
plt.show()
print daily_returns['SPY'].kurtosis()
def assess_portfolio(sd = dt.datetime(2008,1,1), ed = dt.datetime(2009,1,1), \
syms = ['GOOG','AAPL','GLD','XOM'], \
allocs=[0.1,0.2,0.3,0.4], \
sv=1000000, rfr=0.0, sf=252.0, \
gen_plot=True):
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
# Get daily portfolio value
alloc_val = (prices/prices.iloc[0])*allocs*sv
port_val = alloc_val.sum(1)
# Get portfolio statistics (note: std_daily_ret = volatility)
cr = (port_val[-1]-port_val[0])/port_val[0]
adr = cr/sf
sddr = np.std(port_val/port_val[0])
sr = (cr-rfr)/sddr# add code here to compute stats
# Compare daily portfolio value with SPY using a normalized plot
norm_SPY = (prices_SPY/prices_SPY.iloc[0])*sv
if gen_plot:
df_temp = pd.concat([port_val, norm_SPY], keys=['Portfolio', 'SPY'], axis=1)
plot_data(df_temp)
pass
# Add code here to properly compute end value
ev = port_val[-1]
return cr, adr, sddr, sr, ev
def test_run():
# Read data
dates = pd.date_range('2009-01-01', '2012-12-31') # one month only
symbols = ['SPY','XOM','GLD']
df = get_data(symbols, dates)
plot_data(df)
# Compute daily returns
daily_returns = compute_daily_returns(df)
#plot_data(daily_returns, title="Daily returns", ylabel="Daily returns")
# Scatterplot SPY vs XOM
daily_returns.plot(kind='scatter',x='SPY',y='XOM')
beta_XOM,alpha_XOM=np.polyfit(daily_returns['SPY'],daily_returns['XOM'],1)
print "beta_XOM= ",beta_XOM
print "alpha_XOM= ",alpha_XOM
plt.plot(daily_returns['SPY'],beta_XOM*daily_returns['SPY']+alpha_XOM,'-',color='r')
plt.grid()
plt.show()
# Scatterplot SPY vs GLD
daily_returns.plot(kind='scatter',x='SPY',y='GLD')
beta_GLD,alpha_GLD=np.polyfit(daily_returns['SPY'],daily_returns['GLD'],1)
print "beta_GLD= ",beta_GLD
print "alpha_GLD= ",alpha_GLD
plt.plot(daily_returns['SPY'],beta_GLD*daily_returns['SPY']+alpha_GLD,'-',color='r')
plt.grid()
plt.show()
# Calculate correlation coefficient
print daily_returns.corr(method='pearson')
def save_plot_and_model(self,
benchmark_values,
portfolio_value,
trades,
save_plot=False):
"""
save trained model and plot result if save_plot set to True
"""
# save transaction and portfolio image and training records
benchmark_values = benchmark_values.rename(
columns={'SPY': "benchmark"})
portfolio_value = portfolio_value.rename(
columns={'SPY': "q-learn-trader"})
if save_plot:
p_value_all = portfolio_value.join(benchmark_values)
ut.plot_data(trades,
title="transactions_train",
ylabel="amount",
save_image=True,
save_dir=self.current_working_dir)
ut.plot_data(p_value_all,
title="portfolio value_train",
ylabel="USD",
save_image=True,
save_dir=self.current_working_dir)
self.learner.save_model(table_name=self.current_working_dir +
dyna_q_trader_model_save_name)
def calculateMetrics(data):
#accuracy = Mean Absolute % Error MAPE
#efficiency = time taken for training
#25 training data sets
eff = np.zeros([25,2])
accu = np.zeros([25,2])
curr = 0
#set 1-5 : split 90:10 , shuffle 4 times
RunMetric(data,0.9,0,eff,accu)
# set 6-10 : split 80:20 , shuffle 4 times
RunMetric(data, 0.8, 5, eff, accu)
# set 11-15 : split 70:30 , shuffle 4 times
RunMetric(data, 0.7, 10, eff, accu)
# set 16-20 : split 60:40 , shuffle 4 times
RunMetric(data, 0.6, 15, eff, accu)
# set 21-25 : split 50:50 , shuffle 4 times
RunMetric(data, 0.8, 20, eff, accu)
dfeff = pd.DataFrame(eff, columns=['DTLearner', 'RTLearner'], index=range(1,26))
print (dfeff)
util.plot_data(dfeff, title="Train Time - DTLearner vs RTLearner", xlabel="Training sets", ylabel="Time Taken in sec")
dfacc = pd.DataFrame(accu, columns=['DTLearner', 'RTLearner'], index=range(1, 26))
print (dfacc)
util.plot_data(dfacc, title="MAPE - DTLearner vs RTLearner", xlabel="Training sets",
ylabel="MAPE")
def optimize_portfolio(start_date, end_date, symbols):
"""Simulate and optimize portfolio allocations."""
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(start_date, end_date)
prices_all = get_data(symbols, dates) # automatically adds SPY
prices = prices_all[symbols] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
# Get optimal allocations
allocs = find_optimal_allocations(prices)
allocs = allocs / np.sum(allocs) # normalize allocations, if they don't sum to 1.0
# Get daily portfolio value (already normalized since we use default start_val=1.0)
port_val = get_portfolio_value(prices, allocs)
# Get portfolio statistics (note: std_daily_ret = volatility)
cum_ret, avg_daily_ret, std_daily_ret, sharpe_ratio = get_portfolio_stats(port_val)
# Print statistics
print "Start Date:", start_date
print "End Date:", end_date
print "Symbols:", symbols
print "Optimal allocations:", allocs
print "Sharpe Ratio:", sharpe_ratio
print "Volatility (stdev of daily returns):", std_daily_ret
print "Average Daily Return:", avg_daily_ret
print "Cumulative Return:", cum_ret
# Compare daily portfolio value with normalized SPY
normed_SPY = prices_SPY / prices_SPY.ix[0, :]
df_temp = pd.concat([port_val, normed_SPY], keys=['Portfolio', 'SPY'], axis=1)
plot_data(df_temp, title="Daily Portfolio Value and SPY")
def assess_portfolio(sd = dt.datetime(2008,1,1), ed = dt.datetime(2009,1,1), \
syms = ['GOOG','AAPL','GLD','XOM'], \
allocs=[0.1,0.2,0.3,0.4], \
sv=1000000, rfr=0.0, sf=252.0, \
gen_plot=False):
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
# Get daily portfolio value
#port_val = prices_SPY # add code here to compute daily portfolio values
priceSPY=prices_SPY.values
priceSPY /= priceSPY[0]
price_stocks = prices.values
price_stocks /= price_stocks[0]
price_stocks *= allocs
port_val = pd.DataFrame(price_stocks.sum(axis=1),index=prices.index)
prices_SPY = pd.DataFrame(priceSPY,index=prices.index)
# Get portfolio statistics (note: std_daily_ret = volatility)
cr, adr, sddr, sr = [0.25, 0.001, 0.0005, 2.1] # add code here to compute stats
cr = port_val.values[-1] -1
dr = port_val.values
drShift = np.vstack([dr[0],dr[0:(len(dr)-1)]])
dr = dr/drShift -1
adr = dr.mean()
sddr=dr.std()
k = math.sqrt(sf)
sr = k*np.mean(dr-rfr)/np.std(dr-rfr)
# Compare daily portfolio value with SPY using a normalized plot
if gen_plot:
# add code to plot here
df_temp = pd.concat([port_val, prices_SPY], keys=['Portfolio', 'SPY'], axis=1)
df_temp.columns=df_temp.columns.get_level_values(0)
plot_data(df_temp, "Daily portfolio value and SPY", "Date", "Normalized prices")
# Add code here to properly compute end value
ev = sv*port_val.values[-1]
return cr, adr, sddr, sr, ev
def plot_normalized_data(df, title="Normalized prices", xlabel="Date", ylabel="Normalized price"):
"""Normalize given stock prices and plot for comparison.
Parameters
----------
df: DataFrame containing stock prices to plot (non-normalized)
title: plot title
xlabel: X-axis label
ylabel: Y-axis label
"""
normdf = df.divide(df.ix[0])
plot_data(normdf, title=title, xlabel=xlabel, ylabel=ylabel)
def plot_normalized_data(df, title="Normalized prices", xlabel="Date", ylabel="Normalized price"):
"""Normalize given stock prices and plot for comparison.
Parameters
----------
df: DataFrame containing stock prices to plot (non-normalized)
title: plot title
xlabel: X-axis label
ylabel: Y-axis label
"""
#TODO: Your code here
plot_data((df / df.ix[0,:]), title, xlabel,ylabel)
def main():
dates = pd.date_range('2009-01-01', '2012-12-31')
symbols = ['SPY', 'XOM', 'GLD']
df = get_data(symbols, dates)
plot_data(df)
daily_returns = compute_daily_returns(df)
daily_returns.plot(kind='scatter', x='SPY', y='XOM')
beta_XOM, alpha_XOM = np.polyfit(daily_returns['SPY'], daily_returns['XOM'], 1)
plt.show()
daily_returns.plot(kind='scatter', x='SPY', y='GLD')
plt.show()
def plot_normalized_data(df, title="Normalized prices", xlabel="Date", ylabel="Normalized price"):
"""Normalize given stock prices and plot for comparison.
Parameters
----------
df: DataFrame containing stock prices to plot (non-normalized)
title: plot title
xlabel: X-axis label
ylabel: Y-axis label
"""
#TODO: Your code here
# Normalise the data frame
df_temp = df / df.iloc[0]
# plot the normalized data
plot_data(df_temp)
def test_code():
# this is a helper function you can use to test your code
# note that during autograding his function will not be called.
# Define input parameters
of = "./orders/orders.csv"
sv = 10000
# Process orders
portvals = compute_portvals(orders_file=of, start_val=sv)
if isinstance(portvals, pd.DataFrame):
portvals = portvals[portvals.columns[0]] # just get the first column
else:
"warning, code did not return a DataFrame"
# Get portfolio stats
# Here we just fake the data. you should use your code from previous assignments.
#start_dates = dt.datetime(2008, 1, 1)
#end_dates = dt.datetime(2008, 6, 1)
#cum_ret, avg_daily_ret, std_daily_ret, sharpe_ratio = [0.2, 0.01, 0.02, 1.5]
#cum_ret_SPY, avg_daily_ret_SPY, std_daily_ret_SPY, sharpe_ratio_SPY = [0.2, 0.01, 0.02, 1.5]
cum_ret, avg_daily_ret, std_daily_ret, sharpe_ratio, portfolio_value, start_dates, end_dates = compute_portfolio_stats(portvals)
cum_ret_SPY, avg_daily_ret_SPY, std_daily_ret_SPY, sharpe_ratio_SPY, spy_value, start_dates, end_dates = compute_portfolio_stats(get_data(['$SPX'], dates = pd.date_range(start_dates, end_dates)))
portfolio_value = portfolio_value.to_frame()
portfolio_value.columns=[['Portfolio']]
vals = pd.concat([portfolio_value/portfolio_value.iloc[0], spy_value/spy_value.iloc[0]], axis = 1)
plot_data(vals)
# Compare portfolio against $SPX
print
print "Date Range: {} to {}".format(start_dates, end_dates)
print
print "Sharpe Ratio of Fund: {}".format(sharpe_ratio)
print "Sharpe Ratio of $SPX : {}".format(sharpe_ratio_SPY)
print
print "Cumulative Return of Fund: {}".format(cum_ret)
print "Cumulative Return of $SPX : {}".format(cum_ret_SPY)
print
print "Standard Deviation of Fund: {}".format(std_daily_ret)
print "Standard Deviation of $SPX : {}".format(std_daily_ret_SPY)
print
print "Average Daily Return of Fund: {}".format(avg_daily_ret)
print "Average Daily Return of $SPX : {}".format(avg_daily_ret_SPY)
print
print "Final Portfolio Value: {}".format(portvals[-1])
def test_run():
# Read data
dates = pd.date_range('2009-01-01', '2012-12-31')
symbols = ['SPY', 'XOM']
df = get_data(symbols, dates)
plot_data(df)
# Compute daily returns
daily_returns = compute_daily_returns(df)
plot_data(daily_returns, title="Daily returns", ylabel="Daily returns")
# Plot histogram directly from dataframe
daily_returns['SPY'].hist(bins=20, label="SPY")
daily_returns['XOM'].hist(bins=20, label="XOM")
plt.legend(loc='upper right')
plt.show()
def test_run():
#Testing
testing_set = data_set.ix[test_dates]
testX = testing_set.iloc[:, 1:4].as_matrix()
testY = testing_set['Prediction'].as_matrix()
predY = learner.query( testX ) # get the predictions
#Build dataframe to show results
results = pd.DataFrame( predY, columns = ['PredictedY'], index = test_dates )
results['RealizedY'] = testY
results['Error'] = (testY - predY)
rmse = np.sqrt(((testY - predY) ** 2).sum()/testY.shape[0])
plot_data(results[ ['PredictedY', 'RealizedY'] ] , title="Realized vs Predicted", xlabel="Date", ylabel="Price", filename=None)
plot_data( results['Error'], title="Prediction error", xlabel="Date", ylabel="Error", filename=None)
print rmse
def assess_portfolio(sd = dt.datetime(2008,1,1), ed = dt.datetime(2009,1,1), \
syms = ['GOOG','AAPL','GLD','XOM'], \
allocs=[0.1,0.2,0.3,0.4], \
sv=1000000, rfr=0.0, sf=252.0, \
gen_plot=False):
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
# Get daily portfolio value
#port_val = prices_SPY # add code here to compute daily portfolio values
norm=prices/prices.ix[0,:]
norm_allocation=allocs*norm
actual_val=sv*norm_allocation
port_val=actual_val.sum(axis=1)
# Get portfolio statistics (note: std_daily_ret = volatility)
#cr, adr, sddr, sr = [0.25, 0.001, 0.0005, 2.1] # add code here to compute stats
daily_returns = (port_val / port_val.shift(1)) - 1
adr = daily_returns.mean()
cr = (port_val[-1]/port_val[0]) - 1
sddr = daily_returns.std()
k = np.sqrt(sf)
sr = k * np.mean(adr - rfr)/sddr
# Compare daily portfolio value with SPY using a normalized plot
if gen_plot:
# add code to plot here
df_temp = pd.concat([port_val, prices_SPY], keys=['Portfolio', 'SPY'], axis=1)
plot_data(df_temp, title="Daily portfolio value and SPY", xlabel="Date", ylabel="Normalized price")
pass
# Add code here to properly compute end value
ev = port_val[ed]
return cr, adr, sddr, sr, ev
def assess_portfolio(sd = dt.datetime(2008,1,1), ed = dt.datetime(2009,1,1), \
syms = ['GOOG','AAPL','GLD','XOM'], \
allocs=[0.1,0.2,0.3,0.4], \
sv=1000000, rfr=0.0, sf=252.0, \
gen_plot=False):
# Read in adjusted closing prices for given symbols, date range
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
# Get portfolio statistics (note: std_daily_ret = volatility)
cr, adr, sddr, sr, port_val = compute_portfolio_stats(prices, sv, allocs, rfr, sf)
# Compare daily portfolio value with SPY using a normalized plot
if gen_plot:
# add code to plot here
df_temp = pd.concat([port_val/port_val[0], prices_SPY/prices_SPY[0]], keys=['Portfolio', 'SPY'], axis=1)
plot_data(df_temp, ylabel = "Normalized Price")
pass
return cr, adr, sddr, sr, sv
def assess_portfolio(sd = dt.datetime(2008,1,1), ed = dt.datetime(2009,1,1), \
syms = ['GOOG','AAPL','GLD','XOM'], \
allocs = [0.1,0.2,0.3,0.4], \
sv = 1000000, rfr = 0.0, sf = 252.0, \
gen_plot = True):
# Read in adjusted closing prices for given symbols + date range
dates = pd.date_range(sd, ed)
prices_all = get_data(syms, dates) # automatically adds SPY
prices = prices_all[syms] # only portfolio symbols
prices_SPY = prices_all['SPY'] # only SPY, for comparison later
# Get daily portfolio value
prices_SPY = prices_SPY / prices_SPY[0]
normed_prices = prices / prices.ix[0]
alloced = normed_prices * allocs # allocating weights accordingly
print "Allocation to each stock in portfolio over each day is:"
print (alloced)
init_port_distrib = sv * alloced
port_vals = init_port_distrib.sum(axis=1)
# Get portfolio statistics
cr, adr, sddr, sr = compute_portfolio_stats(prices, allocs, rfr, sf)
# Compare daily portfolio value with SPY using a normalized plot
if gen_plot:
df_temp = pd.concat([port_vals/sv, prices_SPY], keys=['Portfolio', 'SPY'], axis=1)
plot_data(df_temp, ylabel="Normalized Price")
plt.show()
pass
# Compute end value
ev = sv * (cr + 1)
return cr, adr, sddr, sr, ev
rmse = math.sqrt(((testY - testPredY) ** 2).sum() / testY.shape[0])
print
print "Out of sample results"
print "RMSE: ", rmse
c = np.corrcoef(testPredY, y=testY)
print "corr: ", c[0,1]
testPredY_pd = pd.DataFrame(data = testPredY, index=testY_pd.index, columns=[symbol])
plotTestY =testPredY_pd*testPrices_pd + testPrices_pd
plotTestPredY = testPredY_pd*testPrices_pd + testPrices_pd
testorders, testlong_orders, testshort_orders, testexit_orders = find_orders(testPredY_pd, 0.025)
testportval = compute_portvals(testorders, start_val=10000)
print "Returns: ", testportval.iloc[-1]/testportval.iloc[0]
plotTestPredData = pd.concat([testPrices_pd, plotTrainY, plotTrainPredY], axis=1)
plotTestPredData.columns = ['Training Prices', 'TrainY', 'PredY']
#TODO If you want to plot...
# plot_data_2axes(plotTrainPredData, ['Predicted Y', 'Training Y'])
plot_data(plotTrainPredData, title='Training Y/Price/Predicted Y:') #Prediction Plot
# plot_data(plotTestPredData)
# plot_corr(testY, testPredY)
plot_orders(trainPrices_pd, trainlong_orders, trainshort_orders, trainexit_orders, title='In Sample Trading Data') #Entries/Exits In-Sample
plot_orders(trainportval/trainportval[0], trainlong_orders, trainshort_orders, trainexit_orders, title='In Sample Backtest', ylabel="Cumulative Return") #In Sample Backtest
plot_orders(testPrices_pd, testlong_orders, testshort_orders, testexit_orders, title='Out of Sample Trading Data') #Entries/Exits Out-Of-Sample
plot_orders(testportval/testportval[0], testlong_orders, testshort_orders, testexit_orders, title='Out of Sample Backtest', ylabel="Cumulative Return") #Out of Sample Backtest
print 'Debug Here'
def test_run():
"""Driver function."""
# Define input parameters
start_date = "2007-12-31"
end_date = "2015-12-31"
# Get quotations
stock_symbol = ["IBM"]
dates = pd.date_range(start_date, end_date)
stock_prices = get_data(stock_symbol, dates, addSPY=False)
stock_prices.dropna(inplace=True)
# Learning and test set dates
# Get only valid dates with non NaN values
n_dates = stock_prices.shape[0]
learning_dates = stock_prices.index.values[0 : int(n_dates * 0.60)]
test_dates = stock_prices.index.values[int(n_dates * 0.60) :]
# print stock_prices.ix[learning_dates,]
print "Learning set from ", learning_dates[0], " to ", learning_dates[-1]
print "Test set from ", test_dates[0], " to ", test_dates[-1]
# Get indicators
bollinger_ind = Bollinger.Bollinger()
momentum_ind = Momentum.Momentum()
volatility_ind = Volatility.Volatility()
bollinger_ind.addPriceSeries(stock_prices)
momentum_ind.addPriceSeries(stock_prices)
volatility_ind.addPriceSeries(stock_prices)
# Get data to be predicted
future_return = stock_prices / stock_prices.shift(5) - 1
future_return.columns = ["Prediction"]
# Merge in a dataframe, combining the three predictors
data_set = future_return.join(bollinger_ind.getIndicator(), how="inner")
data_set = data_set.join(momentum_ind.getIndicator(), how="inner")
data_set = data_set.join(volatility_ind.getIndicator(), how="inner")
data_set.dropna(inplace=True)
# print learning_set
# Learning
learner = BagLearner.BagLearner()
learning_set = data_set.ix[learning_dates]
trainX = learning_set.iloc[:, 1:4].as_matrix()
trainY = learning_set["Prediction"].as_matrix()
learner.addEvidence(trainX, trainY)
# Testing
testing_set = data_set.ix[test_dates]
testX = testing_set.iloc[:, 1:4].as_matrix()
testY = testing_set["Prediction"].as_matrix()
predY = learner.query(testX) # get the predictions
# Build dataframe to show results
results = pd.DataFrame(predY, columns=["PredictedY"], index=test_dates)
results["RealizedY"] = testY
results["Error"] = testY - predY
rmse = np.sqrt(((testY - predY) ** 2).sum() / testY.shape[0])
plot_data(
results[["PredictedY", "RealizedY"]],
title="Realized vs Predicted",
xlabel="Date",
ylabel="Price",
filename=None,
)
plot_data(results["Error"], title="Prediction error", xlabel="Date", ylabel="Error", filename=None)
print rmse
def plot_normalized_data(df, title="Daily portfolio value and SPY", xlabel="Date", ylabel="Normalized price"):
"""Plot stock prices with a custom title and meaningful axis labels."""
df_temp = pd.concat([df, prices_SPY/sv], keys=['Portfolio', 'SPY'], axis=1)
plot_data(df_temp, title, xlabel, ylabel)
def test_run():
# Read data and plot
dates = pd.date_range('2009-01-01', '2012-12-31')
symbols = ['SPY', 'XOM', 'GLD', 'GOOG']
initialize = 'n' # toggles whether or not to initialize with SPY data (must be 'n' if SPY is in symbols)
df = get_data(symbols, dates, initialize)
plot_data(df)
# Compute daily returns
daily_returns = compute_daily_returns(df)
plot_data(daily_returns, title='Daily Returns', xlabel='Date', ylabel='Daily Returns')
# Plot a histogram
daily_returns.hist(bins=20)
# Get mean and stdev
mean = daily_returns['SPY'].mean()
print 'mean=', mean
std = daily_returns['SPY'].std()
print 'stdev=', std
# Scatterplot SPY vs GOOG
daily_returns.plot(kind='scatter', title='Correlation SPY vs GOOG', x='SPY', y='GOOG')
# beta is the slope of the line in the scatterplot, and alpha is the intercept
beta_GOOG, alpha_GOOG = np.polyfit(daily_returns['SPY'], daily_returns['GOOG'], 1) # this returns the alpha and beta for GOOG
# below '-' indicates a line should be drawn
plt.plot(daily_returns['SPY'], beta_GOOG*daily_returns['SPY'] + alpha_GOOG, '-', color='r') # plots a regression curve using y = mx + b
print '\nGOOG beta = ', beta_GOOG
print 'GOOG alpha = ', alpha_GOOG # higher alpha would mean the stock performed better than the SPY
plt.show()
# Scatterplot SPY vs XOM
daily_returns.plot(kind='scatter', title='Correlation SPY vs XOM', x='SPY', y='XOM')
# beta is the slope of the line in the scatterplot, and alpha is the intercept
beta_XOM, alpha_XOM = np.polyfit(daily_returns['SPY'], daily_returns['XOM'], 1) # this returns the alpha and beta for XOM
# below '-' indicates a line should be drawn
plt.plot(daily_returns['SPY'], beta_GOOG*daily_returns['SPY'] + alpha_XOM, '-', color='r') # plots a regression curve using y = mx + b
print '\nXOM beta = ', beta_XOM
print 'XOM alpha = ', alpha_XOM # higher alpha would mean the stock performed better than the SPY
plt.show()
# Calculate Correlation Coefficient
# this uses pearson correlation coeff method to estimate how closely the data points fit the regression line
print '\nCorrelation Coefficients:\n', daily_returns.corr(method='pearson')
# plot separate plots for each stock
plt.axvline(mean, color='w', linestyle='dashed', linewidth=2) # adds vertical dashed line at the mean
plt.axvline(std, color='r', linestyle='dashed', linewidth=2)
plt.axvline(-std, color='r', linestyle='dashed', linewidth=2)
plt.show()
# plot overlayed plots
daily_returns['SPY'].hist(bins=20, label='SPY')
daily_returns['XOM'].hist(bins=20, label='XOM')
daily_returns['GLD'].hist(bins=20, label='GLD')
plt.legend(loc='upper right')
plt.show()
# Compute kurtosis
# This is how closely a dataset fits a gaussian normal distribution +num has longer tails, -num shorter tails
print '\nKurtosis:\n', daily_returns.kurtosis()
