Where is added value in data science?

When I just started my career in data science everyone was into Kaggle competitions, where you were given a data set and your task was to predict the target variable as accurate as possible. Although it is great approach to learn data science, it has very little to do with day-to-day job of data scientists. And moreover it is not where the real value of data science is. First, in business environment your solution don't have to be extremely accurate, it should be reliable and transparent. Second, there many ways of achieving high accuracy without manual stacking and blending of dozens of algorithms (for example, you can use DataRobot or other AutoML solutions). The real value of a data scientist in business setting is to translate business problem into the one than can be solved with 1Gb of data and Random Forest and translate back the feature importance to the relevant KPIs. Of course, you need profound technical skills to do so, but they wouldn't bring much value without soft skills and business acumen. #datascience #ai #addedvalue #carreradvice #datasciencecarrer