A Databricks-based workshop on time series, statistical, and ML forecasting.
Presented to: Adidas North America Data Community
Date: June 2025
- π£ Forecasting foundations: when and why we forecast
- π Time series 101: seasonality, trend, noise
- π Model walkthroughs:
- Naive, Moving Average
- Holt-Winters
- Prophet (stat + ML hybrid)
- XGBoost (pure ML)
- π Model evaluation: MAE, RMSE, prediction intervals
- π§ Business framing: how to choose the βrightβ model
- β Slides + interactive quiz
notebooks/: Databricks-formatted forecasting demoslides/: Full PowerPoint from the training sessionimages/: Visual previews of models (add your screenshots here)data/: (Optional) Sample data for testing or reuse
- Python, SQL
- Databricks Notebooks
- Facebook Prophet, XGBoost
- PowerPoint
- Business analysts learning forecasting
- Data scientists explaining models to stakeholders
- Teams exploring ML vs statistical forecasting tradeoffs