This is the physical activity of determining which is the best strategy to deploy and industrialize the steps previously developed. Here, we will have to consider the best data architecture to be built to automatically acquire data from previously identified data sources, and process them employing the developed data preparation codes. Finally, we will have to define how to feed the model with the processed data, and how to communicate results to the end user.

All of this will have to be documented and audited, in order to ensure a proper level of quality and ease future activities and costs of maintenance.

It should also be mentioned that, especially within more academic and theoretical environments, the deployment plan development will not involve the development of a proper industrial process like the one described above, which will require us to clearly define and formalize, in a sequential way, all performed steps in order to assure replicability of the process, and quality of the obtained results.