A toolbox for a Data Science, most of tools are tailormade for my needs and works with other scripts. The good point that these tools shows how the magic happens, on the other hand, maybe these tools will need an adjust for your needs.
Sometimes, or most of times, people ask me why a tailormade tool if scikit-learn or scipy (for example) do it better and faster.
Respectfully I do not agree with better. Better is very open and wide, something is good for me and bad for you (and vice-versa). These tools exists and are developed because I want have the control of the flow, have access for some specific point of the pipeline hides or just keep recorded the sequence of some part of the pipeline that functions from libraries does not allow/give.
However I need to say that most of all pre-made tools from other libraries will do it faster. but... who said these tailormade tools will be going to production? These tools does not aims to go for production, at production I recommend to use functions from libraries and create a pipeline, these tools are specific for modeling, tunning, exploring or studying phase of the project.
These taylormade tools are for experiments (remember, we are are KarenaLab), here we are commited to the explore, go deeper in all hidden layers.
If you wish to see these tools working and its advantages, take a look at projects and experiments.