Python for Finance Domain¶

What you'll learn¶

• Core financial math (compound interest, returns)
• Working with price time series
• Risk basics (volatility, drawdown)
• The idea of backtesting a strategy
• The library ecosystem

Python dominates quantitative finance — from research notebooks to trading systems. The core math is simple; the libraries (Pandas, NumPy) handle scale. The stdlib examples here are run-verified; the Pandas ones follow its documented API.

Financial math (stdlib, tested)¶

Financial math (stdlib, tested) in Python for Finance — what it is and when to use it.

The foundations need no libraries. Compound interest and simple returns:

def compound(principal, rate, years, n=1):
    """Future value with interest compounded n times per year."""
    return principal * (1 + rate / n) ** (n * years)

print(round(compound(1000, 0.05, 10), 2))   # 10y at 5%, annual

Output:

1628.89

Period-over-period returns from a price series:

def simple_return(prices):
    return [(prices[i] - prices[i-1]) / prices[i-1] for i in range(1, len(prices))]

print([round(r, 4) for r in simple_return([100, 110, 99])])

Output:

[0.1, -0.1]

The price went up 10% then down 10% — note that leaves you below where you started (0.9 × 1.1 = 0.99), a classic reminder that percentage gains and losses aren't symmetric.

Time series with Pandas¶

Index by date, resample, and compute rolling returns on market data.

Real finance work uses Pandas for time-indexed price data:

import pandas as pd     # pip install pandas

# Load prices with a datetime index
prices = pd.read_csv("prices.csv", parse_dates=["date"], index_col="date")

returns = prices["close"].pct_change()          # daily returns
cumulative = (1 + returns).cumprod()            # growth of $1
rolling_vol = returns.rolling(window=20).std()  # 20-day volatility
sma_50 = prices["close"].rolling(50).mean()     # 50-day moving average

Risk basics¶

Volatility, drawdown, and value-at-risk fundamentals.

A few standard risk measures:

• Volatility — standard deviation of returns; how much they swing. Higher = riskier.
• Maximum drawdown — the largest peak-to-trough drop; the worst loss you'd have endured.
• Sharpe ratio — return per unit of risk (excess return ÷ volatility); the classic risk-adjusted measure.

# Max drawdown from a cumulative-growth series (stdlib logic)
def max_drawdown(cumulative):
    peak = cumulative[0]
    worst = 0.0
    for v in cumulative:
        peak = max(peak, v)
        worst = min(worst, (v - peak) / peak)
    return worst

print(round(max_drawdown([1.0, 1.2, 0.9, 1.1]), 4))   # -0.25 (from 1.2 to 0.9)

The drawdown of -0.25 means the portfolio fell 25% from its peak before recovering — the kind of loss an investor actually feels.

Backtesting¶

Simulate a strategy on historical data before risking real money.

Backtesting simulates a trading strategy on historical data to estimate how it would have performed. The skeleton: for each day, decide a signal from past data only, apply it, and track the resulting returns.

Libraries like backtrader, vectorbt, and zipline provide realistic backtesting engines that help avoid these traps.

The ecosystem¶

pandas, NumPy, statsmodels, and quant libraries.

Practice exercises¶

1. Extend simple_return to compute log returns (ln(p_t / p_{t-1})) and compare to simple returns.
2. Write a function for the Sharpe ratio given a list of returns and a risk-free rate.
3. Implement a simple moving-average crossover signal in pure Python (buy when short MA > long MA).
4. Compute max drawdown for a real price series you load with Pandas.
5. List three ways a backtest can look great but mislead you, and how to guard against each.