Heimdall Forecast
Heimdall Forecast is the no-code time series product on Heimdall by RejiCo — build and deploy forecasting models from gold datasets with datetime and value columns.
| Path | Requirement |
|---|---|
| Your data | Gold with Forecast use case, 30+ rows, parseable datetime |
| Sample | AirPassengers in wizard — no Lake |
Open Forecast in the app
Sidebar → Data Intelligence → Forecast (/forecast)
Forecast vs ML vs Loop
| Forecast | ML | Loop | |
|---|---|---|---|
| Output | Future values over time | Single prediction per row | Predictions that learn from feedback |
| Data shape | Datetime + numeric series | Features + target | Loop schema + live examples |
| Gold | Forecast gold (30+ rows) | ML gold (100+ rows) | Not used today |
Typical path
- CSV with date column → Lake bronze
- Publish gold for Forecast
- Build a forecast
- Deploy → Monitor
Or try AirPassengers sample in ~10 minutes: quick start sample.
FAQ
What columns does Forecast gold need?
A parseable datetime column and a numeric value column. Publish with the Forecast use case — see publish for modeling.
Forecast vs ML — which do I pick?
Use Forecast when the question is "what happens next in time?" Use ML for "given these features, what category or number?" See What is Heimdall?.
Can I skip Lake like the AirPassengers sample?
Yes for AirPassengers only in the wizard. Your own time series data needs Lake → Forecast gold.