Author: Anne-Flore Elard, Jan 6 2021
What is analytics and then what is data science and then what is machine learning (and then …)? There are so many confusions that this post will try to recover some meaning about how analytics relates to data science to machine learning and more. Obviously all truths are subjective here 🙂
When I was a student (don’t ask me when), I spent a lot of my time on models related to decisions – how to think structurally about decision problems. I would give away this learning for nothing else in the world. There were two pillars: statistics and optimization models. Today there is a wild variety of models and fields or subfields
Variety is awesome but it can get tricky. To simplify let’s focus on four major groups: Analytics, Data Science, Machine Learning and Artificial Intelligence.
Analytics can be defined as the systematic computation of data. It can rely on statistics but not always. In practice in the companies I worked for, it meant for one to identify historical patterns to help leaders make decisions, for another reporting, for another automation, for another adhoc business analysis.
Data Science leverages a combination of tools to understand or predict an event (e.g. the spread of a pandemic). It relies on a savvy mix of statistics, computer science, math and business knowledge. In practice, I would underline business because the understanding of an event is also the understanding of the business(es) behind it. Great statistics and computer science may give a great model but not necessarily a useful outcome.
Machine Learning relates to algorithms that improve with new data (as time goes by, the models get exposed to more data, depth or breadth, and get better or evolve). In that sense, they mimic some learning abilities that are seen in humans. That makes Machine Learning a subfield of Artificial Intelligence.
Now, Artificial Intelligence as opposed to human intelligence is the ability of machines to apply what they learnt. It implies that machines doted with artificial intelligence programs can perceive their environment, learn from it, adapt to it and take actions
My take on the ecosystem?
*Hint to Kierkegaard’s Concluding Unscientific Postscript (1846) about the subjectivity of the truth