Should You Be a Data Scientist or a Data Analyst?

Lindsey ParkerLindsey ParkerLast updated

If you’re looking to break into tech, you’ve seen the term “data science” thrown around. It’s no surprise HBR named Data Scientist the “sexiest job of the 21st century”; data is more valuable and more available than ever. Making sense of it though? Still a challenge.

Careers in data are mysterious, but luckily for job seekers, patterns are emerging. If you’re analytical-yet-adventurous and you have a flair for storytelling, this could be the field for you.

Data analysts sift through data looking for patterns and solving business problems. They’re pros at data manipulation and have a knack for separating insight from noise.

Data scientists interpret data as analysts do, but their focus is strategic rather than tactical. They use tools such as predictive modeling, machine learning and statistics to find answers. Their work can seem abstract to non-technical decision makers, so they’re often responsible for communicating the story of the data.

Let’s take a deep dive into each career path and discuss how to prepare for that first interview.

Qualifications & Job Responsibilities

Data Analyst

A data analyst turns data into insight. Analytical skills are key, so a STEM degree is often required. Beyond that, you should have:

  • Competence in a variety of languages such as SQL/R/Python
  • Project management and and agile development experience
  • Advanced skills in Excel/Office
  • Data visualization and reporting skills
  • Intellectual curiosity

The bulk of your day will be spent interpreting data. With less emphasis on predictive analytics or statistics, data analysts work with various databases to answer business questions and create reports.

The below posting for a data analyst at Edmodo lists the following responsibilities:

  • Work with stakeholders in Growth, Marketing and Product to understand data and make decisions
  • Develop standardized dashboards and ad hoc reports to inform decision making
  • Validate data and ensure that we make accurate decisions
  • Design ETLs and data warehouse for efficiency and reusability
  • Oversees the design and structure of all intermediate computations
  • Develops and oversees standard methods for combining multiple data streams, even when the data is incomplete or two streams conflict with each other
  • Peer reviewing the designs and work of other analysts
  • Mentoring junior analysts
  • Creating a culture of best practices and quality

All of this experience will carry over into data science, but scientist’s backgrounds are both broader and deeper.

Data Scientist

Data scientists are super-powered analysts with an advanced skill set. As a data scientist you’ll spend lots of time asking future-oriented questions. To find answers, you’ll run experiments and test cutting-edge tech like machine learning.

Source: StackExchange

Per Glassdoor, common job requirements are:

  • An advanced degree (statistics, mathematics or computer science)*
  • Experience using statistical languages (R, Python, etc.)
  • Experience working with/creating data architectures and working with distributed data tools like Hadoop
  • Knowledge of machine learning techniques (clustering, decision tree learning, artificial neural networks, etc.)
  • Knowledge of statistical techniques and concepts (regression, distributions, statistical testing, etc.)
  • Excellent communication skills. Business acumen required to communicate findings to non-technical stakeholders

*Although a high percentage of data scientists hold advanced degrees (88% per KDnuggets), fields of study vary. Your PhD in zoology isn’t a dealbreaker!

A typical day-in-the-life is hard to pin down. You’ll wear many hats and each industry has its own set of expectations. Heavy into statistics and programming, data scientists are systems builders. One day you may build a predictive model based on an interesting trend. Another day you’ll improve data architecture for the analytics team.

A job posting for a data scientist at Evernote lists the following:

  • Apply your expertise in quantitative analysis, data mining, and data visualization to see beyond the numbers, understand how our users interact and engage with Evernote, and draw actionable insights
  • Partner closely with product, engineering and marketing teams to solve problems and identity trends and opportunities
  • Lead the definition of product and business success metrics, collaborate with working teams to ensure correct data tracking instrumentation for measurement
  • Build reports and dashboards to monitor KPIs (Key Performance Indicators), understand drivers of KPI changes
  • Help set up experiments, analyze, and interpret results
  • Effectively communicate insights and recommendations to stakeholders
  • Design and implement predictive models to improve engagement, conversion, and retention
  • Lead ongoing decisions concerning data collection, test/study design and data analyses

If a life of data wrangling appeals to you, you’ll want to put your best foot forward. These jobs are in high demand!

Next Steps

If data analyst sounds like the right fit:

  • Choose a STEM major in computer science, mathematics, etc.
  • Become a wizard in Excel (your coworkers will appreciate it!)
  • Learn a language. Python, SQL, and R are all great choices.
  • Get some hands-on experience. Join a Meetup, volunteer with an organization like DataKind, or contribute to open-source projects.

If data scientist is more your style:

  • Start with the above list! You’ll need a firm foundation in analysis to start.
  • Consider an advanced degree in something like statistics/computer science.
  • Build a portfolio to showcase your skills. Collect open-source contributions and tidy up your GitHub Pages. Showcase different skill sets - clean up some messy data from, do exploratory analysis of an interesting Kaggle dataset, or try a machine learning project.
  • Get some practical experience working through real-world data science questions sourced from Google, Facebook, and Amazon.
  • Practice summarizing insights for a non-technical audience. Explain your projects to your more accommodating friends and family.

Strengthen Your Application:

Flash forward. You’re a Python expert and you can explain the basics of predictive modeling to your mom. You’re ready to apply for your dream job. But you know lots of other candidates have the same qualifications. What can you do to stand out?

Interview Prep

First, assemble your materials. Tailor your portfolio according to the company and job description. Make sure your GitHub profile/blog features a few relevant projects. Code should be visible and well documented. Bonus points for including a README file that explains setup and summarizes the project! Here are a few tips for putting together a standout portfolio:

  • Include (at least) one data cleaning project and one data storytelling project. These match work you’re likely to do on a daily basis.
  • Spend time on visuals. Any graphical representation of your work must clearly tell the right story.
  • Pick topics you’re interested in rather than what’s easy to put together. Your work will be better - you’ll dig deeper, find more interesting connections, and you’ll have fun summarizing.

Once you’ve got a killer portfolio, the best way to prepare for a technical interview is to practice. We recommend Exponent’s Data Science Interview Course which includes questions from Google, Facebook, and Amazon as well as 1-on-1 coaching. If possible, begin far in advance of any interviews so you have time to get comfortable. And don’t forget to practice explaining your answers - communication is as important as technical competence in data careers. Here are a few extra tips:

  • If you’ve never had a technical interview before, it may help to ease into problem-solving under pressure by entering a group challenge. Kaggle competitions are perfect; you choose what to compete for (from low-pressure knowledge challenges through $1,000,000 prizes.) Meetups/forums are great places to meet potential teammates.
  • Try to simulate real-world conditions. Wear your interview outfit, set a timer and work problems in front of an audience.
  • If you’re lucky enough to have friends in data, have them critique you. You can’t prepare for every possible question, but practice speaking on your feet will go along way!

This goes for any interview but bears repeating. Prepare your outfit and paperwork the night before, get a good night’s sleep, and leave early to account for traffic. If you’ve followed the steps above, you’re well-prepared to rock that interview.

To Sum Things Up:

Both data analysts and data scientists make data actionable and "elegant” but a data scientist is a true scientist in the sense that they ask their own questions, figure out how to find answers, and explain how those answers affect the bottom line. The analyst is a super effective problem-solver, but he/she doesn't need 20 slides to explain themselves to upper management.

Keep in mind that data careers are changing constantly. You probably noticed the overlap between the two. It’s true that data analysis is a great entry into data science. The best way to prepare for a career in either is through practice; with the right practical skills, you can own a career in data. Good luck!

Learn everything you need to ace your data science interviews.

Exponent is the fastest-growing tech interview prep platform. Get free interview guides, insider tips, and courses.

Create your free account