Time Series Analysis with Python

Module 1: Introduction to Time Series Analysis

⏳ 1 hour

• Topics: Time series definitions, components, and common use cases.

• Hands-on: Load sample datasets and plot basic time series charts.

Module 2: Data Manipulation & Preprocessing

⏳ 1.5 hours

• Topics: Date/time indexing, resampling, rolling/window functions, handling missing timestamps.

• Hands-on: Clean and resample real-world time series data using Pandas.

Module 3: Visualization & Exploratory Analysis

⏳ 1 hour

• Topics: Line plots, autocorrelation (ACF), partial autocorrelation (PACF), seasonal plots.

• Hands-on: Generate ACF/PACF charts and seasonal subplots to detect patterns.

Module 4: Decomposition & Stationarity Testing

⏳ 1.5 hours

• Topics: Additive vs. multiplicative decomposition, Dickey–Fuller test, detrending techniques.

• Hands-on: Decompose a series and apply stationarity tests; transform data for modeling.

Module 5: Classical Forecasting Models (ARIMA/SARIMA)

⏳ 1.5 hours

• Topics: AR, MA, ARMA, ARIMA, SARIMA concepts, identifying p/d/q hyperparameters.

• Hands-on: Fit and forecast ARIMA/SARIMA models; use grid search for order selection.

Module 6: Exponential Smoothing & Prophet

⏳ 1 hour

• Topics: Simple, Holt’s, and Holt–Winters exponential smoothing; Facebook Prophet basics.

• Hands-on: Build and compare ETS and Prophet forecasts on seasonal data.

Module 7: Model Evaluation & Tuning

⏳ 1 hour

• Topics: Train/test splits for time series, walk-forward validation, error metrics.

• Hands-on: Implement backtesting loops and compute MAE, RMSE, and MAPE for model comparison.

Module 8: Advanced Topics & Deployment

⏳ 1 hour

• Topics: Multivariate forecasting (VAR), intro to LSTM for time series, scheduling forecasts with cron or Airflow.

• Hands-on: Prototype a VAR model and containerize a forecasting script for scheduled runs.

Get certificate