Mariete Forecast Foundation
Forecast is a foundation-model time-series forecaster that lives on a visual canvas. You drag in a data source, drop in covariates, run TimesFM 2.5, and read a forecast that ships with confidence intervals — never a single brittle line. This Foundation course teaches the operator side: how to brief a pipeline, how to read the output, the recipes that pay for themselves in the first month, and how to promote a pipeline to a recurring schedule.
What you'll be able to do.
- Explain forecast is in your own words.
- Building a pipeline
- Reading forecasts
- Workflows and cadence
4 modules. ~60 minutes total.
- 01What Forecast is~12 min
Forecast is not a spreadsheet, not a notebook, and not a chatbot you ask "what will sales be?". Forecast is a time-series forecaster on a vi…
- 02Building a pipeline~14 min
A good Forecast pipeline is built bottom-up, one node at a time, validated at each step. The pattern: > Source · Preprocess · Covariates · F…
- 03Reading forecasts~12 min
When a Forecast pipeline runs, you get a chart, a CSV, and a set of holdout metrics. Reading them well is what separates an operator from a…
- 04Workflows and cadence~12 min
Forecast ships with starter pipelines for the questions teams ask most. They're not separate tools; they're recipes you can run as-is or for…
One sample question. Click to reveal the answer.
Which statement best describes what Forecast is, according to this course?
- A spreadsheet add-in that automates ARIMA modelling
- A foundation-model time-series forecaster on a visual canvas, with confidence intervals as the output
- A chatbot you ask "what will revenue be?" in natural language
- A general-purpose data warehouse with built-in dashboards
15 questions like this. Forecast Foundation pass-mark is 75%.