Python Mean Reversion Alerts: Build an RSI + Bollinger Band Signal Generator

Python Investor’s Toolkit: Part 1: Momentum · Part 2: Monte Carlo · Part 3: Mean Reversion · Part 4: Earnings Quality · Part 5: Macro Regime

Part 3 of 5 — The Python Investor’s Toolkit: Practical quant tools built with free data.

Momentum (Part 1) tells you which stocks are trending. But trends don’t go straight — they oscillate. Mean reversion strategies exploit those oscillations: when a stock gets statistically oversold within an ongoing uptrend, it tends to snap back. That snap-back is where a mean reversion trader earns their edge.

This scanner combines two classic technical indicators — RSI (Relative Strength Index) and Bollinger Bands — to identify stocks that are simultaneously oversold on both measures. A single-indicator signal has too many false positives; the dual confirmation cuts noise significantly.

What You’ll Need

pip install yfinance pandas numpy

How the Two Indicators Work Together

The Code

import yfinance as yf
import pandas as pd
import numpy as np
from datetime import datetime, timedelta

# -------------------------------------------------------
# Indicator functions
# -------------------------------------------------------

def calculate_rsi(prices, period=14):
    """Wilder's RSI — the standard implementation used by most platforms."""
    delta = prices.diff()
    gain  = delta.clip(lower=0)
    loss  = -delta.clip(upper=0)

    # Wilder smoothing (exponential, not simple moving average)
    avg_gain = gain.ewm(alpha=1/period, min_periods=period, adjust=False).mean()
    avg_loss = loss.ewm(alpha=1/period, min_periods=period, adjust=False).mean()

    rs = avg_gain / avg_loss
    return 100 - (100 / (1 + rs))

def calculate_bollinger_bands(prices, period=20, std_dev=2):
    """Standard 20-period Bollinger Bands with 2 standard deviation bands."""
    sma   = prices.rolling(window=period).mean()
    std   = prices.rolling(window=period).std()
    upper = sma + (std * std_dev)
    lower = sma - (std * std_dev)
    return upper, sma, lower

def get_signals(ticker, lookback_days=200):
    """
    Calculate RSI and Bollinger Band signals for a single ticker.
    Returns signal strength: +3 (strong buy) to -3 (strong sell), 0 = neutral.
    """
    end   = datetime.today()
    start = end - timedelta(days=lookback_days + 60)

    data = yf.download(ticker, start=start, end=end, progress=False)
    if data.empty:
        return None

    close = data['Adj Close']
    rsi   = calculate_rsi(close)
    bb_upper, bb_mid, bb_lower = calculate_bollinger_bands(close)

    # %B: where is price within the bands? (0 = lower band, 1 = upper band)
    pct_b = (close - bb_lower) / (bb_upper - bb_lower)

    current_price = float(close.iloc[-1])
    current_rsi   = float(rsi.iloc[-1])
    current_pct_b = float(pct_b.iloc[-1])

    # Classify signal
    if   current_rsi < 30  and current_pct_b < 0.05:  signal, strength = "STRONG BUY",   3
    elif current_rsi < 40  and current_pct_b < 0.20:  signal, strength = "BUY",           2
    elif current_rsi < 45  and current_pct_b < 0.30:  signal, strength = "WEAK BUY",      1
    elif current_rsi > 70  and current_pct_b > 0.95:  signal, strength = "STRONG SELL",  -3
    elif current_rsi > 60  and current_pct_b > 0.80:  signal, strength = "SELL",          -2
    elif current_rsi > 55  and current_pct_b > 0.70:  signal, strength = "WEAK SELL",     -1
    else:                                                signal, strength = "NEUTRAL",       0

    return {
        'ticker'   : ticker,
        'price'    : round(current_price, 2),
        'rsi'      : round(current_rsi, 1),
        'pct_b'    : round(current_pct_b, 3),
        'bb_lower' : round(float(bb_lower.iloc[-1]), 2),
        'bb_upper' : round(float(bb_upper.iloc[-1]), 2),
        'signal'   : signal,
        'strength' : strength
    }

# -------------------------------------------------------
# Scanner: run across a watchlist
# -------------------------------------------------------

def scan_watchlist(tickers):
    """Scan a list of tickers and print actionable signals."""
    print(f"Scanning {len(tickers)} tickers...")
    results = [r for t in tickers if (r := get_signals(t)) is not None]

    df = pd.DataFrame(results).sort_values('strength', ascending=False)

    buys  = df[df['strength'] > 0]
    sells = df[df['strength'] < 0]

    if not buys.empty:
        print("n=== BUY SIGNALS (Oversold) ===")
        print(buys[['ticker','price','rsi','pct_b','signal']].to_string(index=False))

    if not sells.empty:
        print("n=== SELL SIGNALS (Overbought) ===")
        print(sells[['ticker','price','rsi','pct_b','signal']].to_string(index=False))

    if buys.empty and sells.empty:
        print("nNo actionable signals today — all stocks are in neutral territory.")

    df.to_csv('mean_reversion_signals.csv', index=False)
    print(f"nAll results saved to mean_reversion_signals.csv")
    return df

# -------------------------------------------------------
# Run it — replace with your watchlist
# -------------------------------------------------------
watchlist = [
    'AAPL', 'MSFT', 'GOOGL', 'AMZN', 'META', 'NVDA', 'TSLA',
    'JPM', 'BAC', 'GS', 'WMT', 'JNJ', 'XOM', 'CVX',
    'SPY', 'QQQ', 'IWM', 'GLD', 'TLT'
]

results = scan_watchlist(watchlist)

Running It Daily (Automation Tip)

The real value of this tool is consistency — running it every day before market open and acting on signals systematically. On Windows, schedule it with Task Scheduler to run at 8:30 AM. On macOS/Linux, add a cron job:

# Run the scanner Monday–Friday at 8:30 AM
30 8 * * 1-5 /usr/bin/python3 /path/to/mean_reversion_scanner.py >> /path/to/scanner.log 2>&1

What’s Next

Parts 1–3 of this series have focused on price-based signals. Part 4 goes deeper: using SEC EDGAR’s free API to analyze earnings quality directly from financial statements — identifying companies where reported earnings are backed by real cash flow vs. accounting accruals.

Series: The Python Investor’s Toolkit
Part 1: Momentum Scanner
Part 2: Monte Carlo Portfolio Stress-Test
Part 3: Mean Reversion Alert System (this post)
Part 4: Earnings Quality Analyzer (SEC EDGAR)
Part 5: Macro Regime Detector (FRED API)