Data Science, what is it?
You’ve probably heard this term thrown around a lot in the last 10 years, and if you’re a nerd, you have probably even thought ‘wow, that is sexy’… no? And you would be correct! Being a self-proclaimed nerd, I think this term and its related meanings are fantastic. It is the present and future right? Over the next 50 years, we will be driving autonomous cars, wars will be fought with robots and heck, your psychologist might even end up being some algorithm too.
But what is it? I thought I knew what data science was until I was given some professionally transformative experiences over the last few months. I was invited to attend a ‘Data Science for Defence Capability’ course run by the Australian Defence Force Academy in November 2021. On this two-week course, I was to learn what data science is enough so I could become conversational in the matter.
I was super excited, as a studious individual, I researched my teachers for the course to learn their backgrounds and what I would learn from each of them individually. At looking over their professional biographies I thought ‘yup, these are certainly data scientists’. They fitted my presupposed definition to the T, one person-built image recognition algorithms, another built rationale for swarm robots, the other was a programming and algorithm expert, and so on. Would you disagree with my presumption?
So anyway, I was sitting there on my very first day and as a typical course would start, the teaching staff asked, ‘data science, what is it?’. All fourteen students, being engaged and excited, gave their opinion as a matter of fact; ‘machine learning’, ‘AI’, ‘neural networks’, we shouted! Seems like reasonable explanations, right? So, you would then be just as surprised as me when the staff said,
‘Well, we don’t think that we are data scientists.’
We were stunned.
The image recognition specialist, now explaining, said ‘well, you are correct, ML, AI and NN is data science, but not one of us build that technology completely.’ For instance, this specialist explained that their role and speciality was in calculus and building the maths model for neural networks, or using pre-created algorithms, however, that is one slice of the cake in making the end NN tech. This specialist works with the programming expert, they work with a data engineer, work with data analysts and so on.
The class was shown this graphic;
The image recognition specialist actually makes up a minute section of the ‘models and algorithms’ component. Without their skill set, the process would be useless, correct! But without all the ‘stuff’ before and after, their skill set would likewise be useless.
So, you tell me, what or who is a data scientist, are they all data scientists who contribute to this process? Does this question even matter? Why do we need to label? Well, I’ll tell you why this is a very important issue to resolve!
This is important because labelling is how we google, and in this modern-day, googling is how we learn. So, if you saw a carrot and labelled it as a tomato, and then went to google to learn about carrots but instead googled tomato’s, you would learn an awful amount about tomato’s, not carrots. Therefore, in your quest for professional development, you must be very careful about what you are intending to google to learn.
It is all well and good to say you wish to be a data scientist, but what does that mean? Do you want to work with image recognition, robotic behaviour, or economic and business optimisation? Do you want to work with theoretical maths, do you want to be programming or more in the data prep or delivery stages?
Now, before I leave you with this article to mull over, I will offer two more tunes of advice.
Think of the T! The T method is breadth and depth of knowledge. If you go across the T, so broad, you are learning a lot of skills, you might learn engineering, narration and maths, but if you go down the T, so deep, you will be specialising in, let’s say, inferential statistical analysis. There is no correct answer, just do you and what you find interesting, but I would suggest going broad before you dive, never know what you might find out.
Lastly, think out about the analytics continuum, and how far up the maturity scale you wish to go, the further up you go means the more studying you will have to undertake, but if you try to do the entirety of the continuum you will be less capable then someone you focused on a single bit (yes, T-model).
In summary, learn about data science and what you actually want to do, because one thing is very different from the other; apply the T model in your progression, and understand the forms of analysis.
Hit me up at firstname.lastname@example.org
Natassha Selvaraj 'A complete data science guide in 2021' (21 July 2021) Towards Data Science URL - https://towardsdatascience.com/a-complete-data-science-roadmap-in-2021-77a15d6be1d9
Harsha Das 'RoadMap to learn Data Science in 2021' (8 July 2021) LinkedIn Articles - URL: https://www.linkedin.com/pulse/roadmap-learn-data-science-2021-harsha-das/?trk=read_related_article-card_title