Time Series & Forecasting
Predicting the future from the past
Trends, seasonality, and forecasting methods — from classical statistics to ML approaches, and how to evaluate them honestly.
Intermediate3 modules · 11 lessons~93 min865 XPAligned with Hyndman's Forecasting: Principles and PracticeStart with: Classical ML & EnsemblesStart with: Probability & Statistics
Progress0/11
Module
Anatomy of a Series
Trend, seasonality, cycles, and the noise in between
Trend, Seasonality & Cycles8 min · 70 XPTrend, Seasonality & Cycles
Stationarity: Ground Rules for Modeling8 min · 75 XPStationarity: Ground Rules for Modeling
Autocorrelation: The Series Talking to Itself9 min · 80 XPAutocorrelation: The Series Talking to Itself
Module
Classical Forecasting
Sixty years of statistical machinery that still pays the bills
Naive Baselines: Beat Them or Go Home7 min · 65 XPNaive Baselines: Beat Them or Go Home
Moving Averages & Exponential Smoothing9 min · 80 XPMoving Averages & Exponential Smoothing
ARIMA, Intuitively10 min · 90 XPARIMA, Intuitively
Prophet & the Automation Era8 min · 75 XPProphet & the Automation Era
Module
ML Forecasting & Evaluation
Features instead of formulas, and evaluation that can't be fooled
Feature Engineering for Time9 min · 85 XPFeature Engineering for Time
When ML Beats ARIMA9 min · 85 XPWhen ML Beats ARIMA
Never Leak the Future8 min · 80 XPNever Leak the Future
Forecast Metrics That Don't Lie8 min · 80 XPForecast Metrics That Don't Lie
Capstone project
Forecast something real
Pick a real time series — weather, stock prices, your screen time — and forecast it two ways (naive baseline vs. a real model). Score both honestly.
Deliverable: A notebook with both forecasts, error metrics, and what made the series hard to predict.
Milestones
Finish every lesson to earn the Time Series & Forecasting Certificate.