Beta Calculation

Usually, the very first step of analyzing a stock is to know how it relates to the benchmark. That’s the Beta. Let’s build a function to calculate it.

Beta Calculation Function

from data_source_lib import *
# Import our data getting library, which I named data_source_lib.py in my folder.

def beta(ticker, bench="SPY", interval="day"):    
    # will return a dataframe     
    betas = []
    volos = []
    data_get = get_stock_data([bench, ticker], freq="daily")  
    data_get = data_get.get_ondemand_data()
    ben_mark = data_get[data_get.Ticker == bench]
    stock = data_get[data_get.Ticker == ticker]
    
    # Get return and put them in a new dataframe
    ben_mark = ben_mark.rename(columns={"Return": bench + "_bench_re", "Ticker": bench}).reset_index()
    stock = stock.rename(columns={"Return": ticker + "_stock_re", "Ticker": ticker}).reset_index()
    
    new = pd.concat([ben_mark, stock], axis=1)
    new = new[[bench, bench + "_bench_re", ticker, ticker + "_stock_re"]]
    new = new.dropna()
    
    # Calculate beta using covariance matrix
    covmat = np.cov(new[bench + "_bench_re"], new[ticker + "_stock_re"])
    
    # Beta calculation
    beta = covmat[0,1] / np.sqrt(covmat[1,1] * covmat[0,0])
    volatility = np.sqrt(covmat[1,1])
    
    betas.append(beta)
    volos.append(volatility)
    
    betas_df = pd.DataFrame(betas)
    betas_df.columns = ["Beta"]
    
    return betas_df, covmat, volos

Example Usage

beta("AAPL")[0]

Finished AAPL

100%|████████████████████████████████████████████████████████████████████████████████████| 2/2 [00:01<00:00,  1.81it/s]

      Beta
  --- ----------
  0   0.559557