This chapter demonstrated the final two steps in doing a data science project: “Results Interpretation” and “Report and Communication”. They are ignored by the most data scientists if the analytical work is not required by the funding body or mandated project initiator. I would suggest that try to write some kind of report or analytical results’ interpretation even for a data experimental project. The report can be brief or completed. However it is important to go through a series thoughts about the results obtained in response to the initial data analytical problem. Asking whether or not that results answer the original question? Are there any limitations on the results? or any angles haven’t been considered? Most people agree that the analytical results of a data science project is not an engineering solution of a problem. It may need multiple rounds of recursive actions. Sometimes, the analytical results is a starting point of another circle or project. With this understanding, a periodical report is even more important.
A report can contain different contents from process summary to particular model explanation and results interpretation. Numbers with some contextual explanation are useful but graphs can speak more than number and text. Therefore a lot of report using graphical dashboard and data visualization tools.
At the end, it is the sense and opinion which data analytical results supported counts. Further more, how these sense and opinion are understood and accepted by other people rather than the data scientists are the goal of a data science project.