Trends versus predictions
Data analysts use data at an aggregate level to find trends and provide recommendations to improve business performance. Data scientists will use data in machine learning models to predict an event typically at a customer level. Data analysts look at the past to find trends while data scientists use the past to make a prediction about the future.
For example, an ecommerce company will run an email campaign to offer discounts incentivizing customers to make a purchase. A data analyst will report email campaign performance as a whole with metrics such as open rates, click rates, and total sales. The company may want to provide personalized offers for each customer using past behavior such as prior purchases and activity on the website to increase the likelihood of a purchase. In this case, a data scientist will build a machine learning model to predict the best offer to send to the customer.
Job responsibilities
A Google search for the difference between a data scientist and data analyst returns this article as the first result that states business acumen and data storytelling are skills required for a data scientist but not a data analyst. This may be true for a junior analyst but an analyst that hopes to progress to a senior analyst also needs business acumen and the ability to tell a story with data. These skills are required for both data analysts and data scientists to progress further in their careers.
Glassdoor’s data scientist job description lists A/B testing as a responsibility but I’m asked to evaluate A/B tests as a data analyst. In companies with both data analysts and data scientists, data scientists will build machine learning models while data analysts cover all other data analytics responsibilities including A/B test evaluation.
How to tell the difference
Data analyst jobs with data scientist titles will have no mention of building machine learning models.
If the job description contains data analyst responsibilities including A/B testing but no modeling requirements then it’s a hybrid data analyst and data scientist role and may have a data scientist or a data analyst job title.
Example 1 — Data scientist title but no modeling required
This Facebook data scientist role has no mention of building models but does list “experiment results” in the second to last “Product Leadership” bullet point.
Example 2 — Data scientist title but mainly a data analyst with some modeling
This data scientist position mentions modeling but only requires “some experience building data science models”. This is a great job for a mid-level data analyst that has experience building models and looking to transition into data science. Since this has a data scientist title but the job responsibilities are not limited to building models this will leverage skills you’ve developed as a data analyst and allow you to develop modeling skills on the job. If you had limited yourself to data analyst positions this may not have come up in your search.
Example 3 — Data scientist title that requires machine learning experience
This data scientist role is looking for a true data scientist to build machine learning models. The responsibilities include applying machine learning to solve problems and experience with different machine learning techniques.
Conclusion
If you’re a data analyst you may think that you should only search for data analyst jobs because you don’t have the data science skills. However, job titles don’t reflect the actual responsibilities. If you’ve developed all the data analyst skills short of building machine learning models search for data scientist jobs that don’t require building models. If you’re a data analyst working on building your data science skills, look at a crossover role with a data scientist title but requires data analyst skills and some modeling experience. This will help you get the data scientist title and gain more modeling experience to transition to a true data scientist position in the future.