@def title = "Reporting" @def hascode = true
@def tags = ["reporting"]
Presenting your (preliminary) results while developing and after the end of a project is not something that should be neglected. We humans are visual creatures and a good deal of our brain is devoted to processing visual information. Hence, putting some effort into digesting your essential conclusions and visualizing them using adequate plotting libraries really pays off.
Data science is as the name suggests based on data which usually is subject to constant change over the course of a project. Thus, setting up a sustainable reporting pipeline that updates any figures, documents, or tables at the press of a button will save you a lot of trouble further down the road. In this lecture we will learn about the following four frameworks that should help you achieve just that.
Reporting is usually just the icing on the cake. Way more time should be spent on proper data management and model development. However, notebooks are useful for
- one-time self-contained use cases because they provide reporting mechanisms
- and fast prototyping because basically every single cell is a breakpoint.
After you are done experimenting you should migrate the functionality developed in a notebook to your regular code base. Strive towards building a sustainable reusable pipeline especially when dealing with bigger projects and collaborating with other people in your team.