When backtesting algorithmic trading strategies or building financial models in Python, the biggest bottleneck is almost always slow, iterative for loops over Pandas DataFrames.
Whether you are running an intraday strategy on index options or scanning hundreds of equities, vectorizing your technical indicators using pandas and numpy can speed up execution by 50x to 100x.
To save developers and quants from rewriting boilerplate code, I built PyQuant Library — a free, open-access collection of 125+ production-ready, vectorized Python scripts with an embedded browser sandbox.
Here is a quick look at how we handle two classic institutional indicators—Intraday VWAP (with daily session resets) and Supertrend—using clean, vectorized Python.
1. Intraday VWAP with Automatic Daily Session Reset
Standard cumulative VWAP calculations break on multi-day intraday data (like 5-minute candles) unless you group by the trading session date. Here is the clean vectorized way to do it:
python import pandas as pd import numpy as np def calculate_intraday_vwap(df: pd.DataFrame) -> pd.DataFrame: """ Expects a DataFrame with DatetimeIndex and 'High', 'Low', 'Close', 'Volume' columns. Resets calculation automatically at the start of each new trading day. """ df = df.copy() # Calculate Typical Price df['Typical_Price'] = (df['High'] + df['Low'] + df['Close']) / 3 df['TP_Volume'] = df['Typical_Price'] * df['Volume'] # Group by trading date for session reset session_date = df.index.date df['Cum_TP_Vol'] = df.groupby(session_date)['TP_Volume'].cumsum() df['Cum_Vol'] = df.groupby(session_date)['Volume'].cumsum() df['VWAP'] = df['Cum_TP_Vol'] / np.where(df['Cum_Vol'] == 0, np.nan, df['Cum_Vol']) return df.drop(columns=['Typical_Price', 'TP_Volume', 'Cum_TP_Vol', 'Cum_Vol']) 
