import pandas as pd
# constants
ROOT_URL = "http://nefin.com.br/Risk%20Factors/"
FILE_EXT = ".xls"
FACTORS_FILES = {
"Mkt": ROOT_URL + "Market_Factor" + FILE_EXT,
"Market": ROOT_URL + "Market_Factor" + FILE_EXT,
"Rm_minus_Rf": ROOT_URL + "Market_Factor" + FILE_EXT,
"SMB": ROOT_URL + "SMB_Factor" + FILE_EXT,
"HML": ROOT_URL + "HML_Factor" + FILE_EXT,
"WML": ROOT_URL + "WML_Factor" + FILE_EXT,
"IML": ROOT_URL + "IML_Factor" + FILE_EXT,
"Rf": ROOT_URL + "Risk_Free" + FILE_EXT,
"Risk_free": ROOT_URL + "Risk_Free" + FILE_EXT,
"Risk Free": ROOT_URL + "Risk_Free" + FILE_EXT,
"Risk free": ROOT_URL + "Risk_Free" + FILE_EXT,
"Risk-free": ROOT_URL + "Risk_Free" + FILE_EXT,
"Risk-Free": ROOT_URL + "Risk_Free" + FILE_EXT,
}
# dict with aggregation name and pandas resampling frequency
RESAMPLE_FREQ = {
"month": "BM",
"monthly": "BM",
"year": "BY",
"yearly": "BY",
}
# Identity aggregations
IDENTITY_AGG = ["day", "daily", None]
# functions
[docs]def get_single_risk_factor(factor, agg=None, agg_func=None):
"""Download a risk factor time series data from NEFIN's website.
Args:
factor (str): Risk factor to download. Options are: 'Market', 'SMB', 'HML', 'WML', 'IML' and 'Rf'.
agg (str, optional): Frequency to aggregate. Either 'year' (or 'yearly') and 'month' (or 'monhtly'). Defaults to None.
agg_func (str, optional): Function to apply at aggreagtion (e.g. 'last' or 'mean'). If it is None, then it defaults to 'last'.
Returns:
pandas.core.dataframe: Pandas dataframe with (aggregated) time series of chosen risk factor
"""
# get url
url = FACTORS_FILES[factor]
# read xls
print(f"Getting data for factor {factor}...")
df = pd.read_excel(url)
# set index as a datetime from columns year, month, day
dt = [f"{y}-{m}-{d}" for y, m, d in zip(df.year, df.month, df.day)]
df["datetime"] = pd.to_datetime(dt, format="%Y-%m-%d")
df.set_index(["datetime"], inplace=True)
df.drop(["year", "month", "day"], axis=1, inplace=True)
# aggregate if desired
if agg not in IDENTITY_AGG:
if agg_func is None:
print(
"\n\nWARNING: aggregation function not provided. Using last() by default\n"
)
agg_func = "last"
# resample
df = df.resample(RESAMPLE_FREQ[agg]).apply(agg_func)
print("Done!")
return df
[docs]def get_risk_factors(factors=None, agg=None, agg_func=None):
"""Download a risk factor time series data from NEFIN's website.
Args:
factors (str or list-like): Risk factors to download. Options are one or more out of: 'Market', 'SMB', 'HML', 'WML', 'IML' and 'Rf'. If 'all' or None, then all factors will be downloaded.
agg (str, optional): Frequency to aggregate. Either 'year' (or 'yearly') and 'month' (or 'monhtly'). Defaults to None.
agg_func (str, optional): Function to apply at aggreagtion (e.g. 'last' or 'mean'). If it is None, then it defaults to 'last'.
Returns:
pandas.core.dataframe: Pandas dataframe with (aggregated) time series of chosen risk factors
"""
# all factors conditions
if (factors is None) or (isinstance(factors, str) and factors == "all"):
factors = ["Market", "SMB", "HML", "WML", "IML", "Rf"]
# if just one, then convert to list (for iteration)
if isinstance(factors, str) and factors in FACTORS_FILES.keys():
factors = [factors]
# holder of partial data frames
list_dfs = list()
# checker for duplicated factors (due to name flexibilities for some factors)
urls_visited = list()
# loop over desired factors
for factor in set(factors): # set to avoid duplicated
if factor in urls_visited:
continue
else:
urls_visited.append(FACTORS_FILES[factor])
list_dfs.append(get_single_risk_factor(factor, agg=agg, agg_func=agg_func))
# concat all
return pd.concat(list_dfs, axis=1, join="outer")