Skip to content

Kaggle and Kaggle Learn, the pros and cons

  • Article
  • Guide

Read the announcement on Facebook

Kaggle is one of the largest online communities for people who work with data, open to any age, profession, or interest. It works as a social hub: members from around the world connect, work together, and publish their projects as web-based notebooks that others can read and comment on. Its best known feature is the library of datasets that anyone can upload or download, and the competitions that companies and researchers host on the site, sometimes with prize money.

Kaggle Learn

Alongside the datasets and competitions, Kaggle runs a set of short introductory courses under the name Kaggle Learn. They cover data science and machine learning, mostly in Python, and are free. The courses are built around writing and running code rather than around theory. You learn how to call a method and see what it produces, more than why the method works or when it should not be used.

That design is a deliberate trade. For someone with no background it removes the usual friction of setting up an environment and lets them do something real on the first day. For someone who wants to understand the mathematics underneath, it is a starting point rather than a destination.

Pros

  • Free to use, both the site and the courses.
  • The interface is intuitive and easy to pick up.
  • A good starting point for beginners in data science.
  • A place to meet and work with other people who do data science.

Cons

  • Less focus on the theory behind the methods being taught.
  • Not the best place to practise real-world applications. Competition datasets arrive already cleaned and framed, which is not how data shows up in most jobs.

Where it fits

The reasonable way to use Kaggle is as an entry point. Kaggle Learn gets a beginner writing code and seeing results, and the community and datasets give them something to practise on. From there, a learner can decide what they want to go deeper on and where to do it, whether that means a textbook, a formal course, or a project built on messy data they collected themselves.

Sources: The Best Free Data Science Courses Nobody is Talking About, Ken Jee (2020); Pros & cons of Kaggle (For Beginner), K. Bhavsar, Kaggle discussions; Is Kaggle Worth It for Data Scientists?, A. Chugh, Built In (2020)

Originally shared on Facebook.

Query console

Connected to dssoc. Type \h for the commands this console understands.