Data Science in business

Data Science is a hot topic in business, it is considered the new way of doing anything with data, to squeeze value. But is it really like that ? Years ago data science was called with different names: numerical optimization if it was something related to engineering, data mining if it was about finding data into databases, statistics if it was the old venerable art of making inference about data and phenomena.

Today data science is all about data visualization, modeling with tons of AI, and machine learning models. It has almost lost contact with the most important business facts. If DS is working in the engineering domain, it must talk, and respect the rules of engineering, for instance looking at exact solutions of static models (think about civil engineering). If DS is working with surveys an old sector loved by statisticians, it must be able to infer from data.

Unfortunately, the current education on DS is missing these old sources of expertise. No DS is really interested in inference, the old, gold rule of statistics. Prove something with a given level of confidence.

If DS are keen only to build machine learning models virtually ignoring the domain, chances are that they will simply fail, or they will not demonstrate anything useful in business.

So yes, DS is fantastic but business comes first.