Being a data scientist is exciting. It is compatible with almost every sector in business, it’s continuously evolving, and storytelling with data is a skill few people will possess in their lifetime. Whether you’re starting data science early in your career or transitioning from other roles, everyone is excited to get their hands on some data and get to work.
I was someone who was a self taught data scientist. I majored in mathematics and statistics but coding and working with data wasn’t something I had experience in until my first job as a business intelligence developer. I put in work on my own learning about coding in Python, understanding different machine learning techniques, creating cool visualizations, and trying to learn what it meant to be a “data scientist”. For all the fun it is to learn all these new things, in the back of my mind I was nervous about all the things I didn’t know. You see videos on LinkedIn of people creating their own programs using computer vision or dynamic plots that are amazing, but can be intimidating at the same time. I didn’t know how to build a self driving car, did that mean I wasn’t ready for a data scientist role?
Being in the data science profession, here are the different things I wish I would have known heading into my first data scientist position.
1. Clear you assumptions
When I started my data science career I thought I was going to be building these unbelievable machine learning pipelines, doing crazy analytics with deep learning, and predicting things that would cut costs by 90%. A lot of the work is exciting, but each business has different needs and sometimes those needs don’t actually need the machine learning aspect of data science. Being a great data scientist is really just being able to solve problems using data. Period. Regardless if you get there with a machine learning model or just being able to improve the data quality people receive, if you’re good at solving problems people will love you.
2. Learn about non-data science technologies
One of my first jobs was doing data management for professional sports and entertainment franchises. I spent months working through databases without really doing any “data science”. That experience really enhanced my experience working with data and sure enough, a lot of the work I do today is helping colleagues re-code their SQL and database management code because of my experience in the space. Just because you have a data science job doesn’t mean you’re just going to be analyzing linear regression p values all day.
Whether its database management, dashboard visualizations, taking large amount of data and figuring out ways to compress it but still contain value are the things that will consistently present themself to you. Some people like to hand those things off to others because they’re not “data science”. Try to work through those problems because the better you are at coding in one aspect will almost always transfer over to other work you have.
3. Someone else’s work doesn’t diminish yours
Don’t feel the pressure that if you’re new to the data science game that others are creating programs that are way more complex than the projects you’re working on and therefore, you’re not doing anything impactful. Thats completely false. You’ll see something on the news about some company doing amazing robotics using AI and can sometimes be discouraging that you’re not teaching a car how to drive or land a space ship on the moon. You don’t have to do that if it’s not in your interests. I felt pressure early on to learn computer vision because thats what I thought data science was and although it wasn’t a complete waste of time, I didn’t really find it that much fun. I’m more interested in tackling questions of why something happens rather than predict what kind of flower classifier I can build. Find the data science stuff that you enjoy doing the most and build yourself up to be the best data scientist in that area.
4. Presentation of results
I always giggle when I see a data science job posting with a requirement saying “Experience with Microsoft Office.” Like, you’re expecting me to analyze millions of records of sensitive information but you don’t think I understand how to create a Word document? But thats not the point they’re trying to make. What they’re really trying to say is “Do you know how to present your findings when you’re done?” You would be amazed with how advanced a lot of these simple programs have become. People don’t think about how amazing a presentation is, but they surely remember when a presentation is terrible. Take some time to refresh yourself about the basics of programs like Word and PowerPoint. It’s a more valuable skill than you think.
5. You won’t know the answer, and thats a good thing
Your first data science job will be a majority of projects where you think “I have no idea how I am going to figure this out” and thats awesome! There’s nothing interesting about doing a project where anyone could tell you the answer. Embrace the challenge of being creative to solve a problem. Want to mix clustering and time series together but have never seen it done before? Perfect, give it a try! Not everything needs to fit into a decision tree for whether you’re going to do supervised or unsupervised, classification or regression. Remember, you’re the data scientist. People are looking to you for data interpretation. Nobody is going to stop you to ask why you didn’t use a different algorithm because thats not their job. When I don’t know how I’m going to solve a problem, thats when I’m most excited because my most creative work will come out.
Nobody forgets their first job, let alone the first data scientist job. To the average person, a database is a bunch of meaningless information that would take years to interpret. When you’re in the position of a data scientist, you have the ability to show your value right away. Face the challenge head on and when you’re able to get creative with a solution, take the leap. The lessons you learn in your first data scientist role will have an impact in your last role.