Big Data Education

I recently read, Big Data Education: 3 Steps Universities must take

Here are the 3 steps listed:

  1. Data Science cannot be an undergraduate degree
  2. A graduate degree should contain math, stats and computer science
  3. Research

Step 2 seems obvious. Math, stats, and computer science are some of the key areas for data science. I would add communication and presentation skills to the list because people with just math, stats, and CS skills are not known to be naturally good communicators. I agree with step 3. More research needs to be done, but most of the research will need to be interdisiplinary. Universities need to put more effort into interdisiplinary research.

Step 1 confused me a bit. The argument was data science has too many necessary skills and an applied focus area. Of course a person cannot learn everything about data science in an undergraduate degree. Earning a computer science degree does not mean you will know everything about computer science. It just means you know the fundamentals about algorithms, architecture, and operating systems. You know enough about computer science to understand the field and learn more as you go. I think 4 years should be enough time to do the same for data science.

What are your thoughts?






4 responses to “Big Data Education”

  1. gadjo95 Avatar

    Yes 4 year is enough time to learn how to be a data scientist. The problem is that the people at this point of time don’t really have a background as we have after a Master or PhD Degree. I can’t really imagine how you can be efficient in mathematics, computer science and statistics in 4 year after high school. I don’t think a lot of people will think that they want to be a data scientist.

    To be a real data scientist is to have a good expertise in these 3 domains.

    Sorry for my english, It’s not my first language

    1. Ryan Swanstrom Avatar

      Thanks for the comment. Data Science maybe requires more than just a fundamental understanding in the 3 domains.

  2. eballen Avatar

    I have an undergraduate degree in neuroscience and a masters in statistics. Neuroscience in many ways has the same kind of multi-disciplinary nature as data science — psych, bio, physics, chem, etc. I ended up with a shallower but broader education than students with more traditional majors. I think the model used there would work will for a data science B.S. — two year spent on background courses (math, compsci, etc.), then two years focused on integrated and applied work (the equivalent of my neurobiology and physiological psych seminars and courses in neuroscience). In my statistics experience, it was generally assumed that students had almost no statistics experience from undergrad, as the thinking is that you need that time to acquire the necessary math background. Comparing the two models, I prefer the former — more breadth and application as an undergrad, helping students get the basics in several areas and figure out where they want to focus more deep technical study in graduate school or industrial experience.

    1. Ryan Swanstrom Avatar

      Thanks for the extensive response. I like your approach of more breadth as an undergraduate.

Leave a Reply