Blog Post 10 Min Read

Sales Forecast

Hai Bao

Data Science Enthusiast • Jul 15, 2026

Building a sales forecast transforms your revenue from an unpredictable guess into a data-driven roadmap. It allows you to anticipate future revenue, prevent cash flow crises, and make confident decisions about hiring, inventory, and budgeting.

Using a dataset provided by Stats Canada, that contains vehicle sales data in Canada, we will attempt to build a forecast, based on this data.

Fig 1.1: Sales forecast in Power BI. Click on the bottom right corner for fullscreen view.

Forecast Updates

Will be automatically updated as new data from StatsCan is made available through StatsCan Web Data Service (WDS) API.

Methodology

  • Used STL Decomposition since there is evidence of a seasonal trend.
  • Extrapolated data to remove outliers (COVID)
  • Used ARIMA to forecast Trend component.
  • Used XGBoost to forecast Residuals component

    • Walkthrough

    For a step-by-step walkthrough of the process, you can find my Kaggle notebook here