Data scientist. It’s probably the most hyped job title around these days, at least in the field of technology. Why? Because:
- you can easily land a job with a cool company like Netflix, Google or Tesla.
- get assigned to work on cutting-edge solutions like self-driving cars or robotics.
- and receive top compensation for your work.
In fact, even entry-level data scientists can expect to earn $95,000 on average as per the latest Glassdoor survey. Whereas more qualified pros can earn well into $150K annually and above.
So all of the above is rather exciting...but you are not a tech person. Do you still have a chance of breaking into data science? Quite likely if you are keen to learn.
Data scientists come from all different backgrounds. Many have Bachelor’s degrees. Some have no degree at all, but plenty of skill training in the field. Others have Masters or Ph.Ds in related fields, such as math, computer science, and statistics. Others actually did undergraduate studies in other subjects. For instance, Doug Cutting, creator of Hadoop framework, has a degree in linguistics.
Breaking into data science is really a matter of the skills you can accumulate either in college or elsewhere.
So what type of skills do you need to pursue a career as a data scientist? Ideally, it should be a mix of hard (tech/programming) and soft skills in the following areas:
There’s no way around it: you’ll have to learn to code. The best programming language to pick up as a data scientist is Python and/or R.
Python is a sort of a swiss-army knife kind of language for data science as it allows you to build up code for a variety of data science programs. Plus, it has an active community, loads of educational resources and tutorials.
If you are just getting started, check out the following courses:
- Introduction to Python: Absolute Beginner by Microsoft
- Programming with Python: Hands-On Introduction for Beginners on Udemy
- Deep Learning Prerequisites: The Numpy Stack in Python on Udemy
Beyond general programming classes, you should consider specific data science classes, because they will go beyond the coding basics and teach you the core aspects of working with data:
- preparing and cleaning your datasets
- building data pipelines
- using linear models
- and so on.
Learning Python and/or R gives you a certain advantage: you also master the basics of statistical analysis and learn how to automate statistics gathering. Still, you need to understand how different statistical methods work to build more accurate algorithms. In particular, you’ll need to master at least several of the following statistical methods:
- Linear and nonlinear regression
- Resampling methods
- Tree-based statistical models
An Introduction to Statistical Learning (free) written by Hastie, Tibshirani, Witten, and James is a good place to start.
The best way to “cement” those coding and statistics knowledge is by putting them to immediate action. Enter Kaggle – a popular data science community with loads of test data sets, tutorials, courses and challenges you can enter to practice your skills.
Start with doing the simplest challengers and reviewing code written by others. Reach out to the community for help and advice. Practice, practice and practice some more!
Let’s keep it real: oftentimes you’ll get stuck and feel absolutely unmotivated to keep pushing forward.
That’s why having a “virtual” support group that will cheer you up, point up some mistakes and direct towards better approaches of doing X or Y, can be super important. The good news is that data scientists tend to gravitate toward one another. So get set up on Slack and join one of the popular data science groups:
- R-Team for Data Analysis: Global channel for learning R.
- Data Quest: Global Slack networking group for data scientists from.
- Kaggle Noobs: Kaggle and Industry Community.
The short answer is no, it isn’t. But also like a lot of tech careers and positions, your degrees only go so far. They will open certain doors, of course. But your abilities to demonstrate strong expertise through actual project work, whether in coursework you have taken, through initiated projects such as those on Kaggle, or through collaboration with other data scientists through communities you join, will be valued even more. Building up a portfolio of simple data science projects can be just as important as a tech degree.
Of course, some employers will want to see at least a Masters, if not a Ph.D. in data science or a related field. But not all. Many would rather take on avid self-learners with the right attitude and passion for work.
The best advice? Take positions that may not be your final goal but that will move you forward on your path. Ultimately, you may just land that dream data science job.