Group Projects: Kaggle Competition

• Find Competition: Kaggle.com/Competitions
  See existing notebooks at Kaggle.com/code

• Enter Group (3-6 people each): EDS232 Kaggle Group Projects Google Sheet (UCSB login required)

• Describe Competition: Copy EDS232 Kaggle Group Project Template Google Doc (UCSB login required)

• Submit Entry
  See How to Enter a Kaggle Competition (using Kernels) | Kaggle - YouTube

• Present March 9th

  • Time (10 min max per group)
    8 min presentation
    2 min questions
    

  • Not every single member has to present, but everyone in the group has to be available for questions and those not presenting preferably answer the questions
    

  • The format is flexible, but should at a minimum include elements described in the EDS232 Kaggle Group Project Template:

    • About
      group members, competition/notebook name

    • Challenge
      What’s the data-driven question being answered? Why is it important?

    • Data
      What data is available? Describe in terms of response and predictor data with subsets for training, (validating) and testing.

    • Evaluation
      How will your submission be evaluated and scored?

    • Methods
      What types of supervised or unsupervised machine learning techniques could be most applicable?

    • Results
      Please include link to any notebooks attempted and link to leaderboard with group name if actual submission made

    • Conclusion
      What are some of the biggest lessons learned going through this process?
      

  • Please add a link to your group project presentation into EDS232 Kaggle Group Projects