What's a typical day for a data scientist? | Coursera Community
Coursera Header

What's a typical day for a data scientist?

  • 17 January 2019
  • 2 replies

Userlevel 7
  • Community Manager
  • 1275 replies
What's a typical work day like for a data scientist?

Knowing what to expect in a typical work day can be really helpful for people who are considering a data science career. Of course, there are different typical days depending on the job.

Here's a short article answering the question: What does a data analyst really do?

What's your job title, and how would you answer this question?

2 replies

Userlevel 5
Badge +5
Short Answer: No Data Scientist has a typical day.

Long Answer: Data Scientists work on projects and will have a general process for solving problems:

  1. Define problem statement
  2. Get data
  3. Clean/preprocess data
  4. Look for patterns in data (exploratory data analysis)
  5. Build a basic Machine Learning model
  6. Evaluate the model (use test data to assess model accuracy)
  7. Tune model (train the model with different hyperparameters to find the model with the best accuracy)
  8. Present results
This is an iterative process and during this process other things may come up which will change the scope of the project. Or you may realise during the project that it is not actually feasible to continue and take up a new project instead.

Data Scientists must have:
  • Python/R programming skills (or some other functional language - if you know one, you can learn others)
  • Understanding of the core mathematics behind the algorithms (Linear Algebra & Calculus)
  • Familiar with common Machine Learning algorithms (e.g. K-nearest-neighbours, perceptron, neural networks, decision trees)
  • Ability to evaluate results of Machine Learning models (e.g. knowing about how to split train/validation/test data, how to select the right metrics)
  • Ability to present complex mathematical concepts to people who do not come from a mathematics background
  • Presentation skills (ability to tell a story with the data)
  • Familiar with plotting packages (e.g. Python's matplotlib, or R's ggplot2 are commonly used)
Even this is just a short overview and a generalised list. Some Data Scientists will specialise in a specific area (such as Deep Learning), and they may need to know additional things not mentioned here.

Also some Data Scientists will be more involved in the development side of things, so they will need to know more about what to do once you have a trained model. How do you deploy that model? How does it fit in with the systems that you already have in place?
Userlevel 7
Wow, @Liz, this is such a great overview. I love that short and long answers. 🙂 Thank you for taking the time to share your experience with the community!

Has anything surprised you about your work as a data scientist? In other words, did you enter your career with certain expectations that turned out not to be true?