Time Series Analysis — A Handwritten Book

What's Inside

• Stationarity — strict vs weak, why it matters, and the Dickey-Fuller test explained
• Autocorrelation and partial autocorrelation — how to read ACF and PACF plots to pick a model
• AR, MA, ARMA — each derived, with worked parameter-estimation examples
• ARIMA and seasonal ARIMA — the full Box-Jenkins methodology, step by step
• Seasonality decomposition — additive vs multiplicative, STL, and when to use what
• Forecasting — one-step vs multi-step, confidence intervals, forecast error metrics
• Exponential smoothing — simple, double, triple (Holt-Winters) with every formula laid out
• Stationarity transformations — differencing, log, Box-Cox — when each one is the right move
• Real-world pitfalls — spurious regression, look-ahead bias, and data leakage in time series

Who It's For

• Data scientists handling forecasting problems at work who want to understand what their tools are doing
• Students in econometrics, finance, or statistics courses needing a reference that goes beyond slides
• Engineers working on demand planning, capacity modeling, or anomaly detection
• Anyone who has used ARIMA without really knowing what ARIMA is

Frequently Asked

• Do I need a stats background? Basic probability and regression are enough. Everything specific to time series is built up from the ground.
• Is there code? No — this is a notes book, not a cookbook. The focus is intuition and derivation, not syntax.
• Updates? Every buyer gets every future revision, no extra charge.
• Refunds? Digital products aren't refundable once downloaded, but if you hit a defect, email us and we'll fix it or refund.