import pandas as pd
import numpy as np
import yfinance as yfUnlocking Your Trading Potential through Data Analytics
As a seasoned quantitative trader who relies on statistical analysis to thrive in the dynamic world of trading, one of the fundamental questions that drives my strategy is: “What should I prioritize in my trades?” While many traders instinctively gravitate towards high-leverage asset classes in pursuit of maximum profit potential, the real question lies in whether this approach truly amplifies expected returns or simply magnifies risks.
Successful trading isn’t just about occasional windfalls; it’s about consistent profitability. Achieving this demands a crystal-clear understanding of both risk exposure and potential returns.
Instead of fixating solely on leverage, the pivotal inquiry for traders like us becomes: “Where do the most promising opportunities lie?” This question is pivotal, holding the key to unlocking sustained success and profitability in trading endeavors.
Timing is undeniably central to trading prowess. Being in the right asset class can tell you your expected returns and knowing how long your trade will last reduces your risk. But discerning the optimal asset class and timing isn’t an endeavor governed by intuition alone. It demands a rigorous interrogation of data and statistical insights.
In this blog post, I’ll dive into a recent project where I meticulously analyzed the distribution of returns across a diverse spectrum of assets, encompassing bond ETFs, currencies, equity indices, and various sectors. This examination spanned daily, weekly, and monthly time frames, aiming to uncover actionable insights that can inform savvy trading decisions.
Join me as we leverage the power of data analytics to discern the asset classes and temporal windows offering the most promising growth potential for traders like yourself.
Exploring the Data
To begin, I gathered historical data for a diverse set of assets, spanning bond ETFs, currencies, equity indices, and sector-specific stocks. I used the yfinance library in Python to download years of daily adjusted close data for these tickers.
start = '1970-04-01'
end = '2023-04-01'asset_classes = {
'Mid-Cap Stocks': ['MDY', 'IJH', 'XMLV'],
'Mega-Cap Stocks': ['SPY', 'QQQ', 'DIA'],
'Technology': ['XLK', 'VGT', 'FTEC'],
'Healthcare': ['XLV', 'VHT', 'IYH'],
'Financials': ['XLF', 'VFH', 'IYF'],
'Bonds': ['AGG', 'TLT', 'BND'],
'Commodities': ['GLD', 'USO', 'DBC'],
'Real Estate': ['VNQ', 'SCHH', 'XLRE']
}The following code will download the data for the selected assets
# downaload the dat adn save into data folder
# for asset_class, tickers in asset_classes.items():
# for ticker in tickers:
# data = yf.download(ticker, start=start, end=end)
# data.to_csv(f'data/{ticker}.csv')
# print(f'{ticker} downloaded')Lets do a quick Walk through of one csv file so you can see what the data looks like and how we want to handle it.
# download BND data into a dataframe
bnd = pd.read_csv('data/BND.csv', index_col='Date', parse_dates=True)bnd = bnd.sort_index(ascending=False)bnd.head()| Open | High | Low | Close | Adj Close | Volume | |
|---|---|---|---|---|---|---|
| Date | ||||||
| 2023-03-31 | 73.589996 | 73.870003 | 73.470001 | 73.830002 | 71.255219 | 4549200 |
| 2023-03-30 | 73.320000 | 73.519997 | 73.320000 | 73.449997 | 70.888474 | 3455300 |
| 2023-03-29 | 73.160004 | 73.419998 | 73.150002 | 73.349998 | 70.791946 | 4231300 |
| 2023-03-28 | 73.260002 | 73.370003 | 73.199997 | 73.269997 | 70.714745 | 3473300 |
| 2023-03-27 | 73.599998 | 73.730003 | 73.400002 | 73.400002 | 70.840202 | 5769800 |
# new columns close to close daily returns , high to low daily returnsm, and open to close daily returns
bnd['close_to_close'] = bnd['Close'].pct_change().fillna(method='bfill')
bnd['daily_high_low'] = bnd['High'] / bnd['Low'] - 1
bnd['daily_open_close'] = bnd['Close'] / bnd['Open'] - 1/var/folders/nq/y2lyg7p15txfm5ksrw6f90340000gn/T/ipykernel_95669/154206485.py:2: FutureWarning: Series.fillna with 'method' is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.
bnd['close_to_close'] = bnd['Close'].pct_change().fillna(method='bfill')bnd.head()| Open | High | Low | Close | Adj Close | Volume | close_to_close | daily_high_low | daily_open_close | |
|---|---|---|---|---|---|---|---|---|---|
| Date | |||||||||
| 2023-03-31 | 73.589996 | 73.870003 | 73.470001 | 73.830002 | 71.255219 | 4549200 | -0.005147 | 0.005444 | 0.003261 |
| 2023-03-30 | 73.320000 | 73.519997 | 73.320000 | 73.449997 | 70.888474 | 3455300 | -0.005147 | 0.002728 | 0.001773 |
| 2023-03-29 | 73.160004 | 73.419998 | 73.150002 | 73.349998 | 70.791946 | 4231300 | -0.001361 | 0.003691 | 0.002597 |
| 2023-03-28 | 73.260002 | 73.370003 | 73.199997 | 73.269997 | 70.714745 | 3473300 | -0.001091 | 0.002322 | 0.000136 |
| 2023-03-27 | 73.599998 | 73.730003 | 73.400002 | 73.400002 | 70.840202 | 5769800 | 0.001774 | 0.004496 | -0.002717 |
# dropo nan values
bnd = bnd.dropna()
from IPython.display import display
# distribution of returns line for close_to_close, daily_high_low, and daily_open_close overlayed on the same plotly plot
import plotly.graph_objects as go
fig = go.Figure()
fig.add_trace(go.Histogram(x=bnd['close_to_close'], histnorm='probability', name='Close to Close'))
fig.add_trace(go.Histogram(x=bnd['daily_high_low'], histnorm='probability', name='Daily High Low'))
fig.add_trace(go.Histogram(x=bnd['daily_open_close'], histnorm='probability', name='Daily Open Close'))
fig.update_traces(opacity=0.65)
fig.update_layout(barmode='overlay')
fig.update_layout(title='Distribution of Returns', xaxis_title='Return', yaxis_title='Density')
display(fig)
Unable to display output for mime type(s): application/vnd.plotly.v1+json# pritn tabel of summary statistics for close_to_close, daily_high_low, and daily_open_close
summary = bnd[['close_to_close', 'daily_high_low', 'daily_open_close']].describe()
print(summary.to_string(float_format='%.4f'))
close_to_close daily_high_low daily_open_close
count 4024.0000 4024.0000 4024.0000
mean 0.0000 0.0032 -0.0000
std 0.0033 0.0050 0.0027
min -0.0405 0.0000 -0.0377
25% -0.0016 0.0016 -0.0011
50% -0.0001 0.0024 0.0000
75% 0.0015 0.0035 0.0010
max 0.0575 0.2075 0.0403# plot using kde
import seaborn as sns
import matplotlib.pyplot as plt
fig, ax = plt.subplots(1, 3, figsize=(15, 5))
sns.kdeplot(bnd['close_to_close'], ax=ax[0])
ax[0].set_title('Close to Close')
ax[0].set_xlabel('Return')
sns.kdeplot(bnd['daily_high_low'], ax=ax[1])
ax[1].set_title('Daily High Low')
ax[1].set_xlabel('Return')
sns.kdeplot(bnd['daily_open_close'], ax=ax[2])
ax[2].set_title('Daily Open Close')
ax[2].set_xlabel('Return')
plt.tight_layout()
plt.show()
Review
As you can see the we are looking at 3 different values close to close, high to high, and low to low. We will be using the close to close values for our analysis. so that we can understand the whole trading day and after hours. Notice how the daily open to close volatility is pretty low. if your day trading then you are constrained by the open to close volatility.
If you have to pay a fee for a round trip trade of even worse if have zero commisino trding and someoneis front runnign your order flow then you are going to have a bad time. the deck is stacked agains you
Lets build out groups of daily returns.
# Function to compute daily returns
def get_daily_returns(tickers):
data = yf.download(tickers, start='2021-01-01', end='2023-12-31')
returns = data['Adj Close'].pct_change().dropna(how='all', axis=1)
return returns
# Function to compute weekly returns
def get_weekly_returns(tickers):
data = yf.download(tickers, start='2021-01-01', end='2023-12-31')
returns = data['Adj Close'].pct_change().dropna(how='all', axis=1)
weekly_returns = returns.resample('W').last().pct_change()
return weekly_returns
# dislay the asset claas distributions for daily returns
for asset_class, tickers in asset_classes.items():
returns = get_daily_returns(tickers)
fig = go.Figure()
for ticker in returns.columns:
fig.add_trace(go.Histogram(x=returns[ticker], histnorm='probability', name=ticker))
fig.update_traces(opacity=0.65)
fig.update_layout(barmode='overlay')
fig.update_layout(title=f'{asset_class} Daily Returns', xaxis_title='Return', yaxis_title='Density')
display(fig)
[*********************100%%**********************] 3 of 3 completedUnable to display output for mime type(s): application/vnd.plotly.v1+json[*********************100%%**********************] 3 of 3 completedUnable to display output for mime type(s): application/vnd.plotly.v1+json[*********************100%%**********************] 3 of 3 completedUnable to display output for mime type(s): application/vnd.plotly.v1+json[*********************100%%**********************] 3 of 3 completedUnable to display output for mime type(s): application/vnd.plotly.v1+json[*********************100%%**********************] 3 of 3 completedUnable to display output for mime type(s): application/vnd.plotly.v1+json[*********************100%%**********************] 3 of 3 completedUnable to display output for mime type(s): application/vnd.plotly.v1+json[*********************100%%**********************] 3 of 3 completedUnable to display output for mime type(s): application/vnd.plotly.v1+json[*********************100%%**********************] 3 of 3 completedUnable to display output for mime type(s): application/vnd.plotly.v1+json# doensload and siplay weekly returns for asset classes
for asset_class, tickers in asset_classes.items():
returns = get_weekly_returns(tickers)
fig = go.Figure()
for ticker in returns.columns:
fig.add_trace(go.Histogram(x=returns[ticker], histnorm='probability', name=ticker))
fig.update_traces(opacity=0.65)
fig.update_layout(barmode='overlay')
fig.update_layout(title=f'{asset_class} Weekly Returns', xaxis_title='Return', yaxis_title='Density')
display(fig)
[*********************100%%**********************] 3 of 3 completedUnable to display output for mime type(s): application/vnd.plotly.v1+json[*********************100%%**********************] 3 of 3 completedUnable to display output for mime type(s): application/vnd.plotly.v1+json[*********************100%%**********************] 3 of 3 completedUnable to display output for mime type(s): application/vnd.plotly.v1+json[*********************100%%**********************] 3 of 3 completedUnable to display output for mime type(s): application/vnd.plotly.v1+json[*********************100%%**********************] 3 of 3 completedUnable to display output for mime type(s): application/vnd.plotly.v1+json[*********************100%%**********************] 3 of 3 completedUnable to display output for mime type(s): application/vnd.plotly.v1+json[*********************100%%**********************] 3 of 3 completedUnable to display output for mime type(s): application/vnd.plotly.v1+json[*********************100%%**********************] 3 of 3 completedUnable to display output for mime type(s): application/vnd.plotly.v1+json# print the summary statistics for daily returns
for asset_class, tickers in asset_classes.items():
returns = get_daily_returns(tickers)
summary = returns.describe()
print(f'{asset_class} Daily Returns')
print(summary.to_string(float_format='%.4f'))
print('\n')[*********************100%%**********************] 3 of 3 completed
[*********************100%%**********************] 3 of 3 completed
[*********************100%%**********************] 3 of 3 completed
[*********************100%%**********************] 3 of 3 completed
[*********************100%%**********************] 3 of 3 completed
[*********************100%%**********************] 3 of 3 completedMid-Cap Stocks Daily Returns
Ticker IJH MDY XMLV
count 752.0000 752.0000 752.0000
mean 0.0004 0.0004 0.0003
std 0.0131 0.0132 0.0097
min -0.0488 -0.0474 -0.0365
25% -0.0083 -0.0083 -0.0056
50% 0.0003 0.0003 0.0004
75% 0.0086 0.0087 0.0059
max 0.0579 0.0580 0.0421
Mega-Cap Stocks Daily Returns
Ticker DIA QQQ SPY
count 752.0000 752.0000 752.0000
mean 0.0004 0.0005 0.0005
std 0.0094 0.0150 0.0111
min -0.0396 -0.0548 -0.0435
25% -0.0047 -0.0079 -0.0056
50% 0.0006 0.0007 0.0005
75% 0.0059 0.0093 0.0070
max 0.0366 0.0738 0.0550
Technology Daily Returns
Ticker FTEC VGT XLK
count 752.0000 752.0000 752.0000
mean 0.0006 0.0006 0.0007
std 0.0158 0.0159 0.0154
min -0.0522 -0.0524 -0.0531
25% -0.0089 -0.0087 -0.0085
50% 0.0007 0.0007 0.0006
75% 0.0103 0.0105 0.0101
max 0.0837 0.0847 0.0822
Healthcare Daily Returns
Ticker IYH VHT XLV
count 752.0000 752.0000 752.0000
mean 0.0005 0.0003 0.0004
std 0.0094 0.0095 0.0092
min -0.0359 -0.0354 -0.0365
25% -0.0057 -0.0059 -0.0053
50% 0.0005 0.0006 0.0007
75% 0.0063 0.0063 0.0060
max 0.0304 0.0294 0.0306
Financials Daily Returns
Ticker IYF VFH XLF
count 752.0000 752.0000 752.0000
mean 0.0005 0.0005 0.0005
std 0.0125 0.0131 0.0127
min -0.0421 -0.0423 -0.0406
25% -0.0066 -0.0072 -0.0071
50% 0.0007 0.0005 0.0004
75% 0.0078 0.0084 0.0082
max 0.0524 0.0522 0.0505
Bonds Daily Returns
Ticker AGG BND TLT
count 752.0000 752.0000 752.0000
mean -0.0001 -0.0001 -0.0005
std 0.0042 0.0042 0.0112
min -0.0164 -0.0162 -0.0342
25% -0.0026 -0.0027 -0.0080
50% -0.0002 -0.0002 -0.0009
75% 0.0023 0.0023 0.0068
max 0.0215 0.0207 0.0385
[*********************100%%**********************] 3 of 3 completed
[*********************100%%**********************] 3 of 3 completedCommodities Daily Returns
Ticker DBC GLD USO
count 752.0000 752.0000 752.0000
mean 0.0007 0.0001 0.0012
std 0.0133 0.0089 0.0226
min -0.0794 -0.0342 -0.1166
25% -0.0064 -0.0049 -0.0119
50% 0.0016 0.0002 0.0022
75% 0.0088 0.0050 0.0157
max 0.0480 0.0321 0.0791
Real Estate Daily Returns
Ticker SCHH VNQ XLRE
count 752.0000 752.0000 752.0000
mean 0.0003 0.0003 0.0004
std 0.0127 0.0128 0.0130
min -0.0495 -0.0500 -0.0482
25% -0.0068 -0.0068 -0.0069
50% 0.0004 0.0002 0.0007
75% 0.0077 0.0075 0.0079
max 0.0721 0.0731 0.0767
# print weekly returns summary statistics
for asset_class, tickers in asset_classes.items():
returns = get_weekly_returns(tickers)
summary = returns.describe()
print(f'{asset_class} Weekly Returns')
print(summary.to_string(float_format='%.4f'))
print('\n')
[*********************100%%**********************] 3 of 3 completed
/Users/fabianlanderos/miniforge3/envs/quarto/lib/python3.12/site-packages/pandas/core/nanops.py:1016: RuntimeWarning:
invalid value encountered in subtract
[*********************100%%**********************] 3 of 3 completed
/Users/fabianlanderos/miniforge3/envs/quarto/lib/python3.12/site-packages/pandas/core/nanops.py:1016: RuntimeWarning:
invalid value encountered in subtract
[*********************100%%**********************] 3 of 3 completed
[*********************100%%**********************] 3 of 3 completed
/Users/fabianlanderos/miniforge3/envs/quarto/lib/python3.12/site-packages/pandas/core/nanops.py:1016: RuntimeWarning:
invalid value encountered in subtract
/Users/fabianlanderos/miniforge3/envs/quarto/lib/python3.12/site-packages/pandas/core/nanops.py:1016: RuntimeWarning:
invalid value encountered in subtract
/Users/fabianlanderos/miniforge3/envs/quarto/lib/python3.12/site-packages/pandas/core/nanops.py:1016: RuntimeWarning:
invalid value encountered in subtract
[*********************100%%**********************] 3 of 3 completed
/Users/fabianlanderos/miniforge3/envs/quarto/lib/python3.12/site-packages/pandas/core/nanops.py:1016: RuntimeWarning:
invalid value encountered in subtract
/Users/fabianlanderos/miniforge3/envs/quarto/lib/python3.12/site-packages/pandas/core/nanops.py:1016: RuntimeWarning:
invalid value encountered in subtract
/Users/fabianlanderos/miniforge3/envs/quarto/lib/python3.12/site-packages/pandas/core/nanops.py:1016: RuntimeWarning:
invalid value encountered in subtract
[*********************100%%**********************] 3 of 3 completed
/Users/fabianlanderos/miniforge3/envs/quarto/lib/python3.12/site-packages/numpy/core/_methods.py:49: RuntimeWarning:
invalid value encountered in reduce
/Users/fabianlanderos/miniforge3/envs/quarto/lib/python3.12/site-packages/numpy/core/_methods.py:49: RuntimeWarning:
invalid value encountered in reduce
[ 0%% ]Mid-Cap Stocks Weekly Returns
Ticker IJH MDY XMLV
count 155.0000 155.0000 155.0000
mean -2.1297 -0.2165 -inf
std 8.4772 16.7934 NaN
min -76.1494 -94.2530 -inf
25% -1.8172 -1.7651 -2.0535
50% -0.9102 -0.8998 -0.9767
75% -0.0152 -0.0004 -0.1608
max 10.8113 147.9246 42.7911
Mega-Cap Stocks Weekly Returns
Ticker DIA QQQ SPY
count 155.0000 155.0000 155.0000
mean -2.9941 -2.0726 inf
std 32.0140 15.9537 NaN
min -362.5438 -187.9069 -48.9800
25% -1.7802 -1.8162 -1.7643
50% -0.8934 -0.9364 -0.9203
75% 0.1761 0.2459 0.2452
max 115.7719 22.0717 inf
Technology Weekly Returns
Ticker FTEC VGT XLK
count 155.0000 155.0000 155.0000
mean -2.2636 -0.6124 -1.6948
std 15.5887 14.4626 9.3812
min -175.2175 -54.7652 -72.3400
25% -1.9415 -1.9347 -2.0022
50% -0.8666 -0.8622 -0.9079
75% 0.1630 0.1330 0.1061
max 30.6340 145.6055 37.9667
Healthcare Weekly Returns
Ticker IYH VHT XLV
count 155.0000 155.0000 155.0000
mean inf -inf -inf
std NaN NaN NaN
min -86.8380 -inf -inf
25% -2.0127 -1.9122 -2.1054
50% -0.8847 -0.9103 -0.9399
75% 0.0490 -0.0246 -0.2164
max inf 15.9478 42.9784
Financials Weekly Returns
Ticker IYF VFH XLF
count 155.0000 155.0000 155.0000
mean inf inf -inf
std NaN NaN NaN
min -45.6038 -95.6377 -inf
25% -1.8626 -1.8731 -1.9917
50% -0.8786 -0.7538 -0.9862
75% 0.1758 0.1959 0.1047
max inf inf 30.6632
Bonds Weekly Returns
Ticker AGG BND TLT
count 155.0000 155.0000 155.0000
mean NaN NaN -1.9506
std NaN NaN 9.8808
min -inf -inf -108.3204
25% -1.9702 -1.8491 -2.3287
50% -0.9871 -0.9613 -1.1617
75% 0.1324 0.2325 -0.1653
max inf inf 22.6153
[*********************100%%**********************] 3 of 3 completed
/Users/fabianlanderos/miniforge3/envs/quarto/lib/python3.12/site-packages/pandas/core/nanops.py:1016: RuntimeWarning:
invalid value encountered in subtract
/Users/fabianlanderos/miniforge3/envs/quarto/lib/python3.12/site-packages/pandas/core/nanops.py:1016: RuntimeWarning:
invalid value encountered in subtract
[*********************100%%**********************] 3 of 3 completedCommodities Weekly Returns
Ticker DBC GLD USO
count 155.0000 155.0000 155.0000
mean inf inf -1.2287
std NaN NaN 12.4012
min -14.5116 -170.6447 -122.8784
25% -1.9079 -2.0345 -1.9272
50% -1.0488 -0.9900 -0.9322
75% 0.1413 -0.1787 0.2720
max inf inf 44.6151
Real Estate Weekly Returns
Ticker SCHH VNQ XLRE
count 155.0000 155.0000 155.0000
mean -0.5032 inf NaN
std 7.0773 NaN NaN
min -26.0579 -120.7272 -inf
25% -1.8332 -1.8952 -1.8371
50% -0.9557 -0.9898 -0.9452
75% 0.1261 -0.0008 0.0789
max 72.9834 inf inf
/Users/fabianlanderos/miniforge3/envs/quarto/lib/python3.12/site-packages/pandas/core/nanops.py:1016: RuntimeWarning:
invalid value encountered in subtract
/Users/fabianlanderos/miniforge3/envs/quarto/lib/python3.12/site-packages/numpy/core/_methods.py:49: RuntimeWarning:
invalid value encountered in reduce