def test_famafrench_test(self):
self.load_data(start_date="2000-01-01")
try:
quandl.ApiConfig.api_key = 'Z2WYzGME3qmxnqQgBcdX'
ust = quandl.get("USTREASURY/YIELD", start_date="2000-01-01")
f = open('../writeup/tables/FFRegressionTables.tex', 'a')
for co in TestGenericFuture.co_tickers.keys(
): #["Corn", "Brent", "Lumber", "Copper", "Gold", "Wheat"]:
try:
commodity = co
expiries = pd.read_excel("../data/" + commodity +
"ContractExpiries.xlsx")
root_ticker = TestGenericFuture.co_tickers[commodity]
generic_futures = [1, 2]
tickers = [root_ticker + str(i) for i in generic_futures]
near_future_data = self.data.loc[:, tickers[0]]
far_future_data = self.data.loc[:, tickers[1]]
near_future_ticker = tickers[0]
far_future_ticker = tickers[1]
near_future = generic_futures[0]
far_future = generic_futures[1]
near_future_expiries = expiries[[
"Date", str(near_future) + "C_Expiry"
]].set_index("Date")
far_future_expiries = expiries[[
"Date", str(far_future) + "C_Expiry"
]].set_index("Date")
near_future = GenericFuture(ticker=near_future_ticker,
number=near_future,
data=near_future_data)
far_future = GenericFuture(ticker=far_future_ticker,
number=far_future,
data=far_future_data)
near_future.clean_data()
far_future.clean_data()
near_future.set_expiries(near_future_expiries)
far_future.set_expiries(far_future_expiries)
fit = near_future.famafrench_test(far_future,
ust["3 MO"] / 100,
log=False,
frequency='M')
print(commodity)
print(fit.summary())
f.write(fit.summary().as_latex())
f.write('''\newline''')
except KeyError:
print("Error in " + co)
pass
f.close()
except Exception as e:
self.fail(e)
def test_fill_data(self):
self.load_data()
try:
future = GenericFuture(ticker="B", number=1)
tickers = self.data.columns[1:]
data = self.data.iloc[:, 0]
ticker = tickers[1]
number = 1
future = GenericFuture(ticker=ticker, number=number)
future.fill_data(data=data)
future.plot()
plt.show()
except Exception as e:
self.fail()
def test_calculate_return(self):
self.load_data()
try:
future = GenericFuture(ticker="B", number=1)
tickers = self.data.columns[1:]
data = self.data.iloc[:, 0]
ticker = tickers[1]
number = 1
future = GenericFuture(ticker=ticker, number=number, data=data)
future.clean_data()
future.to_monthly()
future.calculate_return()
future.returns().plot()
plt.show()
except Exception as e:
self.fail()
def test_add_future(self):
try:
self.load_data()
tickers = self.data.columns
futures = [GenericFuture(ticker=t, number=int(t.split("_B")[1]), data=self.data[t]) for t in tickers]
brent_curve = CommodityCurve(num_instruments=len(futures))
for f in futures:
brent_curve.add_future(f, f.contract_number())
except Exception as e:
self.fail(e)
def test_clean_data(self):
self.load_data()
try:
future = GenericFuture(ticker="B", number=1)
tickers = self.data.columns[1:]
data = self.data.iloc[:, 0]
ticker = tickers[1]
number = 1
future = GenericFuture(ticker=ticker, number=number, data=data)
future.clean_data()
if future.missing_count() == 0:
self.fail()
print("Missing Data: " + str(future.missing_count()))
future.plot()
plt.show()
except Exception as e:
self.fail()
def test_append_future(self):
try:
self.load_data()
tickers = self.data.columns
futures = [GenericFuture(ticker=t, number=int(t.split("_B")[1]), data=self.data[t]) for t in tickers]
brent_curve = CommodityCurve()
for f in futures:
brent_curve.append_future(f)
brent_curve.create_curve()
except Exception as e:
self.fail(e)
def test_plot_latest(self):
try:
self.load_data()
tickers = self.data.columns
futures = [GenericFuture(ticker=t, number=int(t.split("_B")[1]), data=self.data[t]) for t in tickers]
futures.sort()
brent_curve = CommodityCurve(num_instruments=len(futures))
for f in futures:
brent_curve.add_future(f, f.contract_number())
brent_curve.create_curve()
brent_curve.plot_latest()
plt.show()
except Exception as e:
raise e
#f2 = quandl.get(second_future, start_date='2000-01-01', end_date='2020-12-31')
#f3 = quandl.get(third_future, start_date='2000-01-01', end_date='2020-12-31')
if "Settle" in f1.columns:
price_col = "Settle"
else:
price_col = "Close"
try:
header = ticker.split(sep="_")[1]
except IndexError:
header = ticker.split(sep="_")[0]
#first_f = pd.concat([first_f, f1[price_col].rename(header)], axis=1)
#second_f = pd.concat([second_f, f2[price_col].rename(header)], axis=1)
futures[header] = GenericFuture(ticker=header,
number="1",
data=f1[price_col].rename(header))
print(futures)
from statsmodels.tsa.stattools import adfuller
from numpy import log
from statsmodels.graphics.tsaplots import plot_acf, plot_pacf
from pmdarima.arima.utils import ndiffs
from statsmodels.tsa.arima.model import ARIMA
gasoline = futures['RB']
gasoline.clean_data()
gasoline_monthly = gasoline.series().resample('M').last()
adfuller(gasoline_monthly.diff().dropna(), regression="ct")
def test_calculate_basis(self):
self.load_data(start_date="2000-01-01")
try:
all_basis = []
for co in TestGenericFuture.co_tickers.keys():
try:
commodity = co
expiries = pd.read_excel("../data/" + commodity +
"ContractExpiries.xlsx")
root_ticker = TestGenericFuture.co_tickers[commodity]
generic_futures = [1, 2]
tickers = [root_ticker + str(i) for i in generic_futures]
near_future_data = self.data.loc[:, tickers[0]]
far_future_data = self.data.loc[:, tickers[1]]
near_future_ticker = tickers[0]
far_future_ticker = tickers[1]
near_future = generic_futures[0]
far_future = generic_futures[1]
near_future_expiries = expiries[[
"Date", str(near_future) + "C_Expiry"
]].set_index("Date")
far_future_expiries = expiries[[
"Date", str(far_future) + "C_Expiry"
]].set_index("Date")
near_future = GenericFuture(ticker=near_future_ticker,
number=near_future,
data=near_future_data)
far_future = GenericFuture(ticker=far_future_ticker,
number=far_future,
data=far_future_data)
near_future.clean_data()
far_future.clean_data()
near_future.set_expiries(near_future_expiries)
far_future.set_expiries(far_future_expiries)
basis = near_future.calculate_basis(far_future,
log=True,
ffill=True)
all_basis.append(basis)
except KeyError:
pass
all_basis = pd.concat(all_basis, axis=1)
all_basis.to_excel("../data/AllBasisLog.xlsx")
all_basis = all_basis.resample('M', convention='end').last()
summary_description = all_basis.describe().round(2).to_latex()
f = open('../writeup/tables/SummaryStatistics.tex', 'w')
f.write(summary_description)
f.close()
sns.set(rc={"grid.linewidth": 0.6, "figure.figsize": (11.7, 8.27)})
sns.set_style(style='whitegrid')
ax = sns.boxplot(data=all_basis)
ax.set_xticklabels(ax.get_xticklabels(), rotation=45)
ax.get_figure().savefig("../writeup/img/BasisBoxPlot.png")
#from statsmodels.graphics.tsaplots import plot_acf, plot_pacf
#plot_acf(nb.values.squeeze(), lags=40)
# basis.plot()
#lt.show()
#plot_pacf(near_future.series().values.squeeze(), lags=40)
#plt.show()
except Exception as e:
self.fail(e)
def main(**kwargs):
data = load_data(start_date="2000-01-01")
quandl.ApiConfig.api_key = 'Z2WYzGME3qmxnqQgBcdX'
ust = quandl.get("USTREASURY/YIELD", start_date="2000-01-01")
all_basis = list()
f = open('../tex/FFRegressionTables.tex', 'a')
for co in co_tickers.keys():
try:
commodity = co
expiries = pd.read_excel("../data/" + commodity +
"ContractExpiries.xlsx")
root_ticker = co_tickers[commodity]
generic_futures = [1, 2]
tickers = [root_ticker + str(i) for i in generic_futures]
near_future_data = data.loc[:, tickers[0]]
far_future_data = data.loc[:, tickers[1]]
near_future_ticker = tickers[0]
far_future_ticker = tickers[1]
near_future = generic_futures[0]
far_future = generic_futures[1]
near_future_expiries = expiries[[
"Date", str(near_future) + "C_Expiry"
]].set_index("Date")
far_future_expiries = expiries[[
"Date", str(far_future) + "C_Expiry"
]].set_index("Date")
near_future = GenericFuture(ticker=near_future_ticker,
number=near_future,
data=near_future_data)
far_future = GenericFuture(ticker=far_future_ticker,
number=far_future,
data=far_future_data)
near_future.clean_data()
far_future.clean_data()
near_future.set_expiries(near_future_expiries)
far_future.set_expiries(far_future_expiries)
basis = near_future.calculate_basis(far_future)
all_basis.append(basis)
fit = near_future.famafrench_test(far_future,
ust["3 MO"] / 100,
log=False,
frequency='M')
print(commodity)
print(fit.summary())
f.write(fit.summary().as_latex())
f.write(co)
except KeyError:
print("Error in " + co)
pass
f.close()
all_basis = pd.concat(all_basis, axis=1)
all_basis = all_basis.resample('M', convention='end').last()
all_basis.describe().round(2).drop(
index="count").to_latex("../tex/SummaryStatistics.tex")
ax = sns.boxplot(data=all_basis.drop(columns=["CME_RB1", "CME_PL1"]))
ax.get_figure().savefig("../tex/CYDistribution.png")
return
"Gas": "CME_NG",
"Hogs": "CME_LN",
"Coffee": "ICE_KC",
"Cotton": "ICE_CT",
"Cocoa": "ICE_CC",
"Sugar": "ICE_SB",
"HeatOil": "CME_HO",
"Gasoline": "CME_RB"
}
filename = "../data/TestData_Brent.xlsx"
data = pd.read_excel(filename, index_col="Date")
tickers = data.columns
futures = [
GenericFuture(ticker=t, number=int(t.split("_B")[1]), data=data[t])
for t in tickers
]
futures.sort()
brent_curve = CommodityCurve(num_instruments=len(futures))
for f in futures:
brent_curve.add_future(f, f.contract_number())
brent_curve.create_curve()
brent_curve.to_monthly()
brent_curve.plot_latest()
plt.show()
from sklearn.decomposition import PCA
from sklearn.preprocessing import StandardScaler