Metrics for benchmark comparison

import yfinance as yf
import pandas as pd
import numpy as np
import statistics


# Benchmark data
benchmark = "SPY"
bm_data = yf.download(tickers=benchmark, period="61mo", interval="1mo", prepost=True, progress=False)  # bm = benchmark
bm_data = pd.DataFrame.dropna(bm_data)  # Removes rows with NaN, dividend stocks do this for some reason
bm_close = bm_data['Adj Close']


# Ticker data
ticker = "QQQ"
ticker_data = yf.download(tickers=ticker, period="61mo", interval="1mo", prepost=True, progress=False)
ticker_data = pd.DataFrame.dropna(ticker_data)
ticker_close = ticker_data['Adj Close']




# Correlation coefficient calculation
bm_std = np.std(bm_close)
ticker_std = np.std(ticker_close)
cor_cov = ticker_close.cov(bm_close)
correlation = cor_cov/(bm_std*ticker_std)
correlation_rnd = round(correlation, 2)
print(f"Correlation coefficient: {correlation_rnd}")



# Beta calculation
bm_returns = (bm_close.shift(1) / bm_close)
ticker_returns = (ticker_close.shift(1) / ticker_close)
beta_cov = ticker_returns.cov(bm_returns)
beta_var = bm_returns.var()
beta = beta_cov / beta_var
beta_rnd = round(beta, 2)
print(f"Calculated beta: {beta_rnd}")




# Volume data
# Benchmark data
bm_vol_data = yf.download(tickers=benchmark, period="1mo", interval="1d", prepost=True, progress=False)  # bm = benchmark
bm_vol_data = pd.DataFrame.dropna(bm_vol_data)
bm_vol = bm_vol_data['Volume']
bm_avg_daily_vol = statistics.mean(bm_vol)
# Ticker data
ticker_vol_data = yf.download(tickers=ticker, period="1mo", interval="1d", prepost=True, progress=False)  # bm = benchmark
ticker_vol_data = pd.DataFrame.dropna(ticker_vol_data)
ticker_vol = ticker_vol_data['Volume']
ticker_avg_daily_vol = statistics.mean(ticker_vol)

volume_coefficient = ticker_avg_daily_vol/bm_avg_daily_vol
volume_coefficient_rnd = round(volume_coefficient, 2)

print(f"Volume coefficient: {volume_coefficient_rnd}")




# Novel-alpha prerequisites
# benchmark percentage change
bm_df = bm_close.reset_index()  # Sets date as a column, not index, so it can be removed
bm_hist = pd.DataFrame.from_dict(bm_df)
bm_hist.drop("Date", axis=1, inplace=True)  # deleting date column from dataframe
bm_price_1 = bm_hist.iloc[0, 0]  # retrieves first row
bm_price_2 = bm_hist.iloc[-1, 0]  # retrieves last row
bm_change_percent = ((bm_price_2 - bm_price_1)/bm_price_1)*100

# ticker percentage change
ticker_df = ticker_close.reset_index()
ticker_hist = pd.DataFrame.from_dict(ticker_df)
ticker_hist.drop("Date", axis=1, inplace=True)
ticker_price_1 = ticker_hist.iloc[0, 0]
ticker_price_2 = ticker_hist.iloc[-1, 0]
ticker_change_percent = ((ticker_price_2 - ticker_price_1)/ticker_price_1)*100


# Novel-alpha calculation
bm_change_multiplier = (bm_change_percent + 100)/100
ticker_change_multiplier = (ticker_change_percent + 100)/100
novel_alpha = ticker_change_multiplier/bm_change_multiplier
novel_alpha_rnd = round(novel_alpha, 2)
print(f"Novel-alpha: {novel_alpha_rnd}")




# Novel-gamma
# bm gradient
bm_y = np.array(bm_close)
bm_x = np.array(bm_df['Date'])
bm_x = (len(bm_x))  # converting date(time) into integer array so it can be read by pyplot
bm_xvalues = []
i = 0
while i < bm_x:
    i += 1
    bm_xvalues.append(i)  # this loop converts the absolute value of len() into a list of length "len()"
bm_x = np.array(bm_xvalues)  # converts the list into an array for pyplot

bm_ylength = (len(bm_y))
bm_yvalues_pc = []
t = 0
while t < bm_ylength:
    bm_close_addition = bm_hist.iloc[0+t, 0]
    bm_close_addition_pc = ((bm_close_addition - bm_price_1)/bm_price_1)*100
    bm_yvalues_pc.append(bm_close_addition_pc)
    t += 1  # this loop converts magnitude of price change to percentage change so that the gradient comparison is fair

m, b = np.polyfit(bm_x, bm_yvalues_pc, 1)  # Linear regression
bm_gradient = m

# ticker gradient
ticker_y = np.array(ticker_close)
ticker_x = np.array(ticker_df['Date'])
ticker_x = (len(ticker_x))  # converting datetime into integer array so it can be read by pyplot
ticker_xvalues = []
i = 0
while i < ticker_x:
    i += 1
    ticker_xvalues.append(i)
ticker_x = np.array(ticker_xvalues)

ticker_ylength = (len(ticker_y))
ticker_yvalues_pc = []
t = 0
while t < ticker_ylength:
    ticker_close_addition = ticker_hist.iloc[0+t, 0]
    ticker_close_addition_pc = ((ticker_close_addition - ticker_price_1)/ticker_price_1)*100
    ticker_yvalues_pc.append(ticker_close_addition_pc)
    t += 1  # this loop converts magnitude of price change to percentage change so that the gradient comparison is fair

m, b = np.polyfit(ticker_x, ticker_yvalues_pc, 1)  # Linear regression
ticker_gradient = m

# Novel-gamma calculation
novel_gamma = (ticker_gradient - bm_gradient) + 1
novel_gamma_rnd = round(novel_gamma, 2)
print(f"Novel-gamma: {novel_gamma_rnd}")