rolling_sharpe_ratio_df = m.rolling_sharpe_ratio(df_data,
columns_name=columns_name,
window_length=36,
min_periods=36,
start_gap=6,
benchmark=0.02)
### Rolling alpha
rolling_alpha_df = m.rolling_alpha(df_data,
columns_name=columns_name,
window_length=36,
min_periods=36,
start_gap=6)
### Rolling correlation
rolling_corr_df = m.rolling_corr(df_data,
columns_name=columns_name,
target_benchmark='Russell 3000',
window_length=36,
min_periods=36,
start_gap=6)
### Draw Down
dd_df = 100 * m.draw_down(df_data, columns_name, start_gap)
### Graph for result
with PdfPages('Rolling Ratio Figure.pdf') as pdf:
plt.style.use('fivethirtyeight')
rolling_annual_return_df[[
'TeamCo Client Composite', 'HFRI Fund Weighted Composite Index',
'HFRI Fund of Funds Composite Index'
]].plot(title='36 Months Rolling Annual Return')
plt.legend(prop={'size': 12})
pdf.savefig()
window_length,
min_periods, MAR,
threshold)
### Rolling omega ratio
rolling_omega_ratio_df = m.rolling_omega_ratio(df_data, columns_name,
window_length, min_periods,
MAR)
### Rolling sharp ratio
rolling_sharpe_ratio_df = m.rolling_sharpe_ratio(df_data, columns_name,
window_length,
min_periods, benchmark)
### Rolling alpha
rolling_alpha_df = m.rolling_alpha(df_data, columns_name, window_length,
min_periods)
### Rolling correlation
rolling_corr_df = m.rolling_corr(df_data, columns_name, market_index,
window_length, min_periods)
### Draw Down
dd_df = 100 * m.draw_down(df_data, columns_name)
### Generate graph and save them to the pdf file
p.graph_gen('Rolling Ratio Figure and Radar Chart Result.pdf', index_name, rolling_annual_return_df, cum_return_df, \
rolling_alpha_df,rolling_beta_df, rolling_corr_df, rolling_sharpe_ratio_df, \
rolling_sortino_ratio_df,rolling_omega_ratio_df, dd_df, Beta_df, Beta_df_p, \
Beta_df_np, Corr_df, Corr_df_p, Corr_df_np)
### Output all static dataframe into excel file
dfs = [Annulized_Return_df,Calendar_Return_df,Sharpe_df,Sortino_df,\
Standard_deviation_df,Downside_Deviation_df,Beta_df,Beta_df_p,\
Beta_df_np,Omega_df,Corr_df,Corr_df_p,Corr_df_np,Summary_table_df]
put.multiple_dfs(dfs, 'Financial Ratio', 'Financial Ratio Result.xlsx', 1)
window_length,
min_periods, MAR,
threshold)
### Rolling omega ratio
rolling_omega_ratio_df = m.rolling_omega_ratio(df_data, columns_name,
window_length, min_periods,
MAR)
### Rolling sharp ratio
rolling_sharpe_ratio_df = m.rolling_sharpe_ratio(df_data, columns_name,
window_length,
min_periods, benchmark)
### Rolling alpha
rolling_alpha_df = m.rolling_alpha(df_data, columns_name, window_length,
min_periods)
### Rolling correlation
rolling_corr_df = m.rolling_corr(df_data, columns_name, market_index,
window_length_corr, min_periods_corr)
### Calculate the correlation with other fund's mean return
# Modify the market data, replace it with mean of other fund's return
df_data_corr = df_data.copy()
df_data_corr['Russell 3000'] = df_data_other.mean(
axis=1) # replace with population mean
final_cum_return = cum_return_df_other.iloc[-1, :-1]
quantile = pd.qcut(final_cum_return.T, 4, labels=['Q1', 'Q2', 'Q3', 'Q4'])
Q1 = quantile[quantile.values == 'Q1'].index
Q2 = quantile[quantile.values == 'Q2'].index
Q3 = quantile[quantile.values == 'Q3'].index
Q4 = quantile[quantile.values == 'Q4'].index
df_data_Q1 = df_data.copy()
df_data_Q1['Russell 3000'] = df_data_other[Q1].mean(axis=1)
rolling_beta_df = m.rolling_beta(df_data, columns_name, window_length, min_periods) ### Rolling annulized return rolling_annual_return_df = m.rolling_annulized_return(df_data, columns_name, window_length, min_periods) ### Cummulative return cum_return_df = m.cumulative_return(df_data, columns_name, window_length, min_periods) cum_return_df_other = m.cumulative_return(df_data_other, columns_name_other, window_length, min_periods) ### Rolling sortino ratio rolling_sortino_ratio_df = m.rolling_sortino_ratio(df_data, columns_name, window_length, min_periods, MAR, threshold) ### Rolling omega ratio rolling_omega_ratio_df = m.rolling_omega_ratio(df_data, columns_name, window_length, min_periods, MAR) ### Rolling sharp ratio rolling_sharpe_ratio_df = m.rolling_sharpe_ratio(df_data, columns_name, window_length, min_periods, benchmark) ### Rolling alpha rolling_alpha_df = m.rolling_alpha(df_data, columns_name, window_length, min_periods) ### Rolling correlation rolling_corr_df = m.rolling_corr(df_data, columns_name, market_index, window_length_corr, min_periods_corr) ### Calculate the correlation with other fund's mean return # Modify the market data, replace it with mean of other fund's return df_data_corr = df_data.copy() df_data_corr['Russell 3000'] = df_data_other.mean(axis=1) # replace with population mean final_cum_return = cum_return_df_other.iloc[-1,:-1] quantile = pd.qcut(final_cum_return.T,4, labels=['Q1','Q2','Q3','Q4']) Q1 = quantile[quantile.values == 'Q1'].index Q2 = quantile[quantile.values == 'Q2'].index Q3 = quantile[quantile.values == 'Q3'].index Q4 = quantile[quantile.values == 'Q4'].index df_data_Q1 = df_data.copy() df_data_Q1['Russell 3000'] = df_data_other[Q1].mean(axis=1) df_data_Q2 = df_data.copy()