Data science as a term was coined back in 1996 in an obscure journal.
Later in 2012, D.J. Patil and Thomas Davenport wrote an article in HBR titled: "Data Scientist: The Sexiest Job of the 21st Century" popularizing the term.
This new dream job, combined the coding skills of a programmer with a statistician's ability to build mathematical models to create tremendous value for companies.
The only problem at the time was the lack of a curriculum. If you wanted to get into the field, you had to figure things out on your own. You had to deal with a lot of ambiguity and uncertainty.
Those who possessed this rare and valuable combination of skills could write their own ticket to riches. They had a career moat.
As demand for data scientists has increased, online courses, books, certifications and even college degrees have sprung up. You no longer have to figure things out on your own.
But something else happened also.
The skills are still valuable but not as rare. With the increasing availability of remote work, companies can hire cheaper talent from across the world. That initial moat is slowly disappearing.
So how do you ensure that you build a lasting career moat?
The secret is unstructured learning!
Whether you’re trying to explore something new in your field, or master an existing skill, unstructured learning is the key to moving quickly.
Unstructured learning is a type of learning without a curriculum. You could put together a curriculum based on college courses, if the topic exists, but if it’s new, the best way to do it is to design an interesting project.
After publishing my book on SQL Patterns, I’ve been getting a lot of questions about how to study advanced SQL. People want to know if there are any books or courses available.
But that’s the wrong approach.
Why?
Because we learn best and remember most when we struggle to answer our own questions.
When you do a project, you have to struggle with the exact questions you need to learn vs. what a certain curriculum deems most important. And if your project closely resembles real world ones, even better.
Companies don't really care about what degrees you have, they care about the value you can create for them
It's pretty common these days to graduate college with a masters degree, sometimes even a PhD, and not be able to find a job.
If a company has a choice between a fresh graduate with a masters degree but no experience in the field and a person with no degree but a few years experience, they will hire the person with experience 9 times out of 10.
Experience is a more direct way to demonstrate that you can create value. One way you can get that experience is through unstructured learning by solving real world problems.
You can also do these projects while you have a job. Ask around to see what other people are working on, pick something you like, get permission to do it during work hours, or don’t bother with permission and do it on your own time.
That's why whenever people ask me for advice on how to advance their careers in a new, unfamiliar area, I always suggest they take on a project in that area as part of their job, or even on their own time.
Many lucrative careers require skills you cannot learn in college
There are so many lucrative careers built in the last 5-7 years, with skills you simply can't learn in college (such as podcasting, social media influencer, Facebook marketer, etc.)
Not only is there no curriculum for them, often you can't even learn these things from books.
Maybe becoming a top Instagram influencer isn’t what you want in life, but if you think about how these people learned the skills that made them influential or got them the job, you can't help but realize the power of unstructured learning.
They learned how to build a following and how to create valuable content and they also learned ancillary skills such as video and audio editing, presenting information in interesting ways, speaking in front of a camera or microphone, connecting with others, etc.
Many of these are incredibly valuable, evergreen skills that will serve them for life. If tomorrow one of these platforms disappears, the skills will still be useful and can easily be transferred onto new and as of yet undiscovered platforms.
More importantly they learned how to learn in unstructured ways, the most powerful, evergreen meta skill.
If there's a college curriculum, someone younger, hungrier and cheaper than you can be trained to do your job
I love this tweet by Naval Ravikant:
This quote provides the most compelling reason for unstructured learning.
Because this "specific knowledge" that Naval is referring to cannot be trained for, you have to learn it yourself, by solving real world problems through projects.
Combine this with learning in public and you can increase your luck surface area and attract opportunities you wouldn’t have known about otherwise.
I’ve done this myself several times and landed some really cool gigs.
SQL Pattern of the Week (SQLPOW)
This week’s pattern revolves around finding non-matching rows in a table. I had to use it recently to find data that didn’t get processed by an address verification function.
It’s also very useful to use as a way to filter data from a table based on another table, assuming you can find a common key between them.
There are two ways to do this. They both can be performant on different databases. I once changed pattern 2 to pattern 1 decreasing a query’s execution time from an hour to 30 seconds in an older MySQL instance without breaking the business logic.
Pattern 1: Correlated subquery
SELECT a.* FROM a WHERE NOT EXISTS (SELECT * FROM b WHERE a.key = b.key)
Pattern 2: Left join with NULL filter
SELECT a.* FROM a LEFT OUTER JOIN b ON a.key = b.key WHERE b.key IS NULL
Until next time!