SuggestGrid has data containers named as types. Types are used to store actions and to return recommendations. For instance you can create a type for rating actions.
A type could be explicit or implicit. This feature must be set for each type. Implicit actions are plain observations of user data where it's not possible to evaluate the feedback as positive or negative. Explicit actions are observations on user actions that could be positive or negative.
Implicit actions are actions that are just observed. They might indicate a positive a negative experience. Some examples of implicit actions are views, purchases, or add to card actions.
Explicit actions also indicate a preference. They must come from explicit intent from users. Actions on ratings, like buttons, or analysed comments with rating scores are some examples of explicit actions.
You can refer to Apache Spark documentation on implicit and explicit actions for more details.
Once you create a type, the second step is to send actions for that type.