Data Science Methodology Approaches

To develop and deploy any, and particularly complex engineering system successfully, a systematic application of the appropriate methodology is crucial. Organizational systems, in which various data analytics models and software services are deployed intensively, are not the exceptions. A goal of this lecture is to raise the level of students’ recognition of the importance and applicability of various data science methodology approaches.
In this lecture, we outline the fundamental notions and a role of information engineering, business intelligence, data science, data analytics, data mining, and data science projects. Then, we give an overview of existing data science methodology frameworks, i.e. process models, and discuss the analogies with software process models based on the life-cycle philosophy. Finally, we give an overview of The CRoss Industry Standard Process for Data Mining (CRISP-DM) process model and discuss the role of all its six phases, with a particular attention to the practice of Business Understanding, Data Understanding, and Data Preparation phases, as they are of a crucial importance to identify the context and setup the scope, goals, and expected values of each data science project.
After the lecture, the students will be able to understand the common terminology, and clearly recognize the importance of a selection, preparation, and application of a proper data science methodology framework in data science projects of diverse sizes and natures. They will be able to understand all the data science methodology steps and recognize CRISP-DM as a common methodology framework with initial ideas how it can be applied in data science practice. The students will become ready for additional study work, including self-study efforts, to learn how to actively apply business understanding, data understanding, and data preparation phases in their data science practice.