Before a pipeline

A prediction pipeline can only use data and setups that already exist. Create those before you add steps.

Before you start

You can open CAPS. The worker URLs for DHIS2, CHAP, and the climate service are already set. That install work is CHAP, climate, and DHIS2.

Checklist

1. Data elements and datasets for climate covariates, thresholds, and prediction results.
2. An approved model evaluation and prediction setup in the Modelling App.
3. Climate datasets imported for the covariates the model uses.
4. A connection to each source DHIS2 instance when the health data lives somewhere other than this instance. Skip this when the modelling instance already holds the case data.

Check

You can name the dataset, the approved evaluation, and the climate datasets the pipeline will use. If the pipeline will pull from another instance, that instance appears on About.

Next

Create a pipeline.