Many people often ask me if a degree in Data Science is worth it. As someone who has a degree and works with some people with degrees and some without, I'll shed some light on if a degree might be right for you.
Data science is one of those fields where many top/visible people we see have a Master’s or PhD in Machine Learning or something similarly analytical.
The short answer is: it depends.
For women and many people of color, it’s often harder to prove our worth. We’re frequently underpaid, undervalued, taken advantage of, and pushed out of many technical roles. I do think academic achievements help recruiters (who can ultimately decide if we even move to a technical conversation during the hiring process) see us closer on par to a white male who may not have a degree. For most of us, this comes at a cost, but there are affordable programs and bootcamps if you don't already have some coding or statistical experience. While a minority of Data Scientist, and really all data oriented tech roles have PhDs, most with ML Engineer and Data Science job titles have at least a Master's degree. If you have a high school diploma it may be easier to Data Engineering from Database Administration or other roles that require specific certifications over hard math and engineering skills.
For some industries, an advanced degree can be essential to landing a high paying role. In the field of Data Science, it's hard to prove yourself when the specialization deals with a lot of gatekeepers looking to spot a "fake" or impostor. Having a degree when you don't already have a technical background can set you apart and provide some validity. Besides, it's not uncommon for highly skilled Data Scientists to have inconsistent Github activity. I was recently asked by the head of AI/ML at my company if it was normal for data scientists, and out of the people I've interviewed their commits and PRs are far fewer and further between than software engineering folks. Note that this is a way to get ahead and set yourself apart.
For those coming from software engineering or business analytics, this may be less of a concern. Typically software engineers can learn statistical modeling without needing a formal education. Similarly, analysts learn more formal object-oriented programming and software engineering techniques.
There is tremendous opportunity to come into Data Science from your domain or the industry you already know about. I got my start by taking what I already knew (marketing) and learning coding and stats to apply it. Out of all the ways to start a career in Data Science / ML the hardest path is from your domain. Typically it means you have to learn both the statistics and coding, but it's not impossible. It may take longer to fully transition, but there are plenty of us with more time than money especially when it comes to taking out loans for school.
This is hard, but I know you're not reading this because you're scared of a challenge. I know a few people that got into this industry without a technical degree, but if that's you, I'd love to hear your story and how you did it. Until then here are my tips as someone who has interviewed people for Data Scientist roles.
You will probably be questioned to death about your hard math skills if you have the formula for Naive Bayes memorized and statistics problems they expect you to do using just the back of a napkin.
While this is often the most difficult to learn there are amazing resources for the beginner, intermediate machine learning math noob. You should be able to tell the difference between descriptive and prescriptive statistics, describe gradient descent, and the basics of regression. If you already know the basics you can get a copy of Math for Machine Learning for FREE from the publisher's website.
The other aspect you'll be grilled on is your coding chops. I recommend learning Python and SQL as they're some of the most popular languages used by Data Scientists Python is like a swiss army knife in that you can perform data analysis, build ML models, and build data pipelines to clean and transform your data. SQL is mostly used to retrieve data from a database and is written using keywords that are close to what they mean in plain English so it's fairly easy to learn.
In some interviews you'll be asked to code live or complete a take home project. Either way, you want to practice and fall in metaphorical love with one scripting language an know it well. My go to is Python, but R and Python (the top languages for data science)
This is the hardest part. It always feels like it's the chicken and the egg scenario, how can you get a job if you can't get experience to get the job. Here you've got to be more creative. Leverage social media and sites like Upwork to start freelancing. When I first got into tech I was an SEO master and would find local black businesses to see if they had a google business page set up. Then I'd work with them for next to nothing except a discount and boost their visibility in search rankings. Truth is, while most companies want to use machine learning, they're ill-prepared to hire someone to do it. A good way to avoid the annoyance of interviewing with these types of shops is to work hourly for businesses with a small budget.
I got so much shit from people when I was a communications major, and yet in my technical role that skill is my best asset. When you're dealing with technical work you need to be able to explain that to non-technical folks around you. This means both verbally and visually. We talk a lot about visual communication, but being concise and being able to discern how much technical jargon you can go into with whom you're talking with.
Most of us can see a well designed visualization, but it's hard to know the factors that separate the good from the bad. I suggest the book Storytelling with Data for anyone who has a handle of excel, but wants to make their charts more readable and Cool Infographics for true newbies who want to see examples of good and badly designed visualizations.
As I've covered in previous blogs, my networking skills came in handy when I started changing careers. I needed to be in touch with a whole new set of people. I didn't know anyone in the industry so I leveraged LinkedIn and Twitter to meet new people. Some of whom have recommended me for jobs and I've found gems on how to grow my skills just from checking the timeline. Start talking to people at meetups and if you aren't in a big city with a plethora of Data Science events to go to, you can find your tribe online.
If you want to talk to me or other badass folks more about careers in Data Science feel free to join my Slack channel for fullyConnected. fullyConnected is my (upcoming) platform to learn data science and machine learning with a focus on ethics and how to use data to improve the lives of marginalized people.
It's up to you to determine if a degree is worth the cost and risk. I'll be honest, I wouldn't have gotten my current job if I did the bare minimum in school and I supplemented my education with a lot of online courses and networking. My degree did get me in the door and I've noticed a pattern that my company frequently hires new grads. At the end of the day, it's really about getting solid experience and how well you can explain your work. There will be plenty of companies who won't bat an eye about rejecting someone with no formal education, but they weren't the right fit for you anyway. They should realize this is no way to build a diverse business, but often they don't care. While we can pressure them to change these outdated attitudes, the best thing you can do is look for companies that align with your values.