I’m passing along an excellent post by Terence Shin that I caught on the data analysis blog KDnuggets. Mr. Shin’s post speaks to those of you who have a great enthusiasm for the latest hot thing in data analysis — the skills or tools that are being discussed on the blogs and in the press. Currently, one of these hot things is machine learning. There is always something.

Read the post yourself. It is very well done. Its key points are: (1) Machine learning (or any hot thing in data analysis) is only one part of the toolkit that you will have to wield to solve real-world problems, and (2) A proper engagement of machine learning (or any advanced skill) requires an understanding of foundational skills.

To my mind, learning data analysis is similar to learning any occupational skill. It is not altogether different from learning to be a craftsperson, like an electrician or carpenter. One group of trainees learns how to wire a house, and another learns to translate raw data into information to support decision-making. It is possible to write and execute scripts to perform highly-complicated statistical operations using documentation and web searches, much like I could perform a do-it-yourself rewiring of my home by watching YouTube videos. However, as you gain practical experience in this field, you will find that there is much more to the job than knowing how to wield the most complicated piece of equipment.

This kind of learning strategy seems likely to hinder students’ overall development as data analysts. Start by mastering learning basic tools to solve simple problems, and work your way up to more complex problems and more sophisticated tools. This strategy puts you in the position of a problem-solver and question-answerer, and something more than a programming tech. I think such a strategy is more likely to result in you being a competent entry-level analyst upon graduation.

That being said, I do not want to discourage students from experimenting with stuff while they are learning. That passion will keep you pushing your boundaries and learning over your career, because there is plenty to learn over a lifetime (I’m still learning). Do not extinguish that passion. Just make sure that you don’t become over-focused on this or that method to the detriment of becoming a solid analyst with a good grasp of their toolkit.

Photo Credit. Johnson, Paula J, and Michael Crummett. Wheelwrights and cartwrights Dale Thibault and Harvey Howes, Miles City, Montana. United States Miles City Montana, 1979. Miles City, Montana. Photograph. https://www.loc.gov/item/afc1981005_01_22928/.

Leave a Reply

Your email address will not be published.