Due to changes in a third-party data source or due to internal data pipeline changes, a model might encounter unexpected types of data. It might receive a “Big int” when it was expecting an int; or a unicode string when it was expecting ASCII. It’s common to also encounter situations when fields that should be ints are interchangeably both ints and strings (as well as vice versa!). All these situations often cause unnecessary processing errors and incorrect predictions.
Learn more about type mismatch and how it affects your model in production.