Our data science team has been engaged for some time productionising a simulation model whose function is to predict demand for acute hospital care. Providing the functionality required to make the model easy to use, reliable, and quick in use has posed a considerable challenge, as has building, maintaining, and keeping updated this model while it is being used by a range of users with different needs.

This talk will explore firstly how the team tackled this with data science methods as well as by using the full range of skills within a multidisciplinary analytic team. Running the model required a range of data science technologies, including cloud compute and storage, Docker, R and Python code, applications, R/ Shiny, and databricks. 

It takes a village to ship a model, and the talk will also describe the work of the multidisciplinary team who worked on the model- working on its design, analytical questions which came from the results and parameters, user researchers, and a team to train and guide users of the model to assist them in high quality modelling to support the model's use in business cases.