Forecasting
and Anomaly Detection

Without any ML expertise required

• Leverage the structure of your Kumo graph to:
  • Effectively incorporate complementary, substitutionary, and other correlation effects between time series in your forecasts
  • Support forecasting on cold start time series with no historical data
• Fluidly transition from forecasting at a company/region level down to the most granular level in your data (e.g. across millions of item per region per day combinations)
• Explore how changes to your planning assumptions (in prices, product assortment, external events, etc.) can change your forecasts
• Easily add 3rd party weather, holidays, events, macroeconomic/demographic data to enrich your predictive accuracy