In the previous chapter, we learned how to define a business problem from a data science perspective through a very structured approach, which included how to identify and understand business requirements, an approach to solutioning it, and how to build data pipelines and carry out analysis.
In this chapter, we will look at the reproducibility of computational work and research practices, which is a major challenge faced today across the industry, as well as by academics—especially in data science work, in which most of the data, complete datasets, and associated workflow cannot be accessed completely.
Today, most research and technical papers conclude with the approach used on the sample data, a brief mention of the methodology used, and a theoretical approach to a solution. Most of these works lack detailed calculations and step-by-step approaches. This is a very limited amount of knowledge for anyone reading it to be able to reproduce the same work that was carried out. This...