Historical beta calculation via Python API

djp

What would be the best way to replicate the historical beta calculation through the Python API (preferably via lseg-data)? As I understand, there is currently no example available in the Workspace Codebook library of such a historical beta calculation, as the API instead only allows for obtaining the latest available data points?

Specifically, I am looking for an efficient way of replicating the below Excel-based template file retrieving Company Historical Beta, provided to me by the support team.

Jirapongse

@djp

Thank you for reaching out to us.

You can check the Calculation sheet that show how to calculate those values. The sample code looks like this:

def calculateBeta(ric, index, end, interval, datapoints, beta=''):
    stock_resp = historical_pricing.summaries.Definition(
        ric, 
        interval=interval,    
        end = end, 
        count = datapoints,
        fields=['TRDPRC_1'],
        extended_params={"includeTradeIndicator":"True"}
    ).get_data()
    stock_df = stock_resp.data.df.ffill()
    if 'TRADE_IND' in stock_df.columns:
            stock_df = stock_df.drop(['TRADE_IND'], axis=1)
    stock_reverse = stock_df.iloc[::-1]
    stock_reverse1 = stock_df.shift()[::-1]
    stock1 = np.log(stock_reverse / stock_reverse1)
    stock1 = stock1.dropna()
    #display(stock1)

    index_resp = historical_pricing.summaries.Definition(
        index, 
        interval=interval,    
        end = end, 
        count = datapoints,
        fields=['TRDPRC_1'],
        extended_params={"includeTradeIndicator":"True"}
    ).get_data()
    index_df = index_resp.data.df.ffill()
    if 'TRADE_IND' in index_df.columns:
            index_df = index_df.drop(['TRADE_IND'], axis=1)
    index_reverse = index_df.iloc[::-1]
    index_reverse1 = index_df.shift()[::-1]
    index1 = np.log(index_reverse / index_reverse1)
    index1 = index1.dropna()

    temp = pd.DataFrame()

    if beta=='up':
        temp['INDEX'] = index1.apply(lambda x: np.nan if x['TRDPRC_1'] <= 0 else x['TRDPRC_1'],axis=1)
        temp['STOCK'] = stock1['TRDPRC_1']
        temp = temp.dropna()
    elif beta=='down':
        temp['INDEX'] = index1.apply(lambda x: np.nan if x['TRDPRC_1'] >= 0 else x['TRDPRC_1'],axis=1)
        temp['STOCK'] = stock1['TRDPRC_1']
        temp = temp.dropna()
    else:
        temp['INDEX'] = index1['TRDPRC_1']
        temp['STOCK'] = stock1['TRDPRC_1']
   
    slope, _ = np.polyfit(temp['INDEX'].tolist(), temp['STOCK'].tolist(), 1)    
    return slope

It can be used like this:

_90_days_beta_daily = calculateBeta("PTT.BK",".SETI","2025-05-20",Intervals.DAILY,91)
_180_days_beta_daily = calculateBeta("PTT.BK",".SETI","2025-05-20",Intervals.DAILY,181)
_2_years_beta_weekly = calculateBeta("PTT.BK",".SETI","2025-05-20",Intervals.WEEKLY,105)
_3_years_beta_weekly = calculateBeta("PTT.BK",".SETI","2025-05-20",Intervals.WEEKLY,157)
_5_years_beta_monthly = calculateBeta("PTT.BK",".SETI","2025-05-20",Intervals.MONTHLY,61)

_2_years_beta_up_weekly = calculateBeta("PTT.BK",".SETI","2025-05-20",Intervals.WEEKLY,105,'up')
_3_years_beta_up_weekly = calculateBeta("PTT.BK",".SETI","2025-05-20",Intervals.WEEKLY,157,'up')
_5_years_beta_up_weekly = calculateBeta("PTT.BK",".SETI","2025-05-20",Intervals.MONTHLY,61,'up')

_2_years_beta_down_weekly = calculateBeta("PTT.BK",".SETI","2025-05-20",Intervals.WEEKLY,105,'down')
_3_years_beta_down_weekly = calculateBeta("PTT.BK",".SETI","2025-05-20",Intervals.WEEKLY,157,'down')
_5_years_beta_down_weekly = calculateBeta("PTT.BK",".SETI","2025-05-20",Intervals.MONTHLY,61,'down')

The code is not fully tested. It may not support all use cases so feel free to modify it according to your requirement.