Accessing and preparing data within KNIME

The stock prediction application consists of three parts: data access, model training, and deployment. The deployment workflow orchestrates the data access and model training steps by calling them from separate workflows. Thus, the same data preparation and model training steps could also be called from other deployment workflows. Figure 14.4 shows the part of the deployment workflow that calls the remote workflows (available on KNIME Hub at https://kni.me/w/4JrfiNV6NrqE7VKo). Further details of the deployment workflow will be introduced in the Consuming the H2O model in the deployment application section:

Figure 14.4 – Calling data preprocessing and forecasting steps from the deployment workflow

The deployment workflow generates the forecasts as follows:

1. The workflow loops over one stock symbol at a time with the Table Row To Variable Loop Start node.
2. Next, it calls the data access and preprocessing...