These articles were controversial in the sense that they highlighted the differences between data science and other disciplines, at a time when many believed that data science was just old stuff being re-branded, or being practiced by people knowing nothing about statistics. Ironically, some of the old stuff actually re-branded itself as data science, not the other way around.
Analytics practitioners and users grew by a factor 5 over the last three years, faster than they can be properly trained, despite the numerous programs available for free, including ours (for self-learners only). Thus many are not equipped with the proper training. This created an opportunity to develop efficient, simple methods that could be understood and properly used by the layman, and even by robots, to process modern, big data. Unfortunately, this aspect of data science is considered by many, even today, to not be part of the core data science framework: it has created much of the controversy, mostly around the concept of automated data science, automated machine learning, or automated statistical science, including the introduction of new powerful algorithms such as automated indexation - a very fast clustering algorithm for big, unstructured text data - to create large taxonomies, by companies such as Amazon or Google.
Here is my selection of controversial articles: