Machine learning codes are usually laden with far too many complexities (hyperparameters, preprocessing and right things coded in wrong places ...). I am on a mission to untangle ML code by logically separating work flow and creating a re-usable framework that you can easily apply to your machine learning tasks involving any type of data.

Machine learning code involves phases of operations such as load, pre-process, visualise, train, cross validate and post-process. It's these phases that I exploit in creating this framework to reduce the complexity involved in machine learning tasks.

Please see this post to see how this framework connects to the overall organisational strategy for data analytics.