Beyond Statistical Modeling

00:08 Introduction
00:30 The majority of us are still getting to grips with classical statistical modelling. What is the key idea today?
03:36 What are Amazon and Google doing better than what other data science teams can do?
06:54 What is so wrong about the Kaggle mindset?
10:14 Why would you say that the majority of data science departments are still stuck in academic ways?
12:07 Changing the entire business model to adapt to a new technology is certainly risky. Is this the reason why there is a resistance to this sort of change?
14:45 People are always talking about buzzwords like AI, Deep Learning, and so on. How much confidence can you have that in ten years time those words will still be a topic of importance?
17:33 What should people do to instill this culture in their data science teams? How should they push their teams to take more risks and be more groundbreaking?
21:01 Is the fact that executives currently do not have the level of mechanical sympathy that they need, that makes them big blockers?


Despite many companies currently investing huge amounts of resources into establishing data science capabilities, the hard truth is that the majority of these projects fail to impact upon daily business operations. As such, we look beyond classic statistical modelling and explain why the impact of a proper data science department must run far deeper than just collecting data, manipulating it and getting an output.

Contrary to popular belief, when it comes to extracting patterns or replicating some form of human intelligence, all we have is statistics. These statistical methods may have names such as ““deep learning”“, or sometimes they’re referred to as ““AI”“, but this doesn’t change the fact that they remain statistical models at their core.

In most companies, data science teams aren’t really doing ““data science”“; more often than not, what they’re doing is statistical modeling. It’s simply not comparable to Microsoft, Google or Amazon where instead of playing with deep learning, they’re already imagining what the next form of deep learning will be.

Data science is often about going beyond the frame that is already in place and trying to project into the future. As a company, you ask yourself, what is missing tech-wise? Innovation usually comes from requirements. It is far too easy to invest in the latest buzzword technology, which usually only brings a cosmetic change, without thinking about how it actually impacts the business.

To wrap things up, we talk more about how many didn’t successfully anticipate the growth of Amazon and Alibaba and we offer a simple litmus test that companies can do to help their data science teams push for innovation.