Kaggle is a platform for data science competitions where participants can learn from experts, work with interesting real-world datasets, and potentially win prizes. Competitions involve building predictive models on a training dataset, submitting predictions for a test dataset, and being scored based on accuracy. Competitors can use any open-source language and should focus on feature engineering, trying different algorithms and models, and building ensembles of multiple models to improve performance. Participating is a way to learn and improve skills in machine learning and data science.

1. How to get started in
Merja Kajava
September 1, 2015
Helsinki Data Analytics and Science Meetup

2. is a competition platform
for (aspiring) data
scientists

4. Why participate in Kaggle
The data
Competition steps
Tips

5. Why participate in
Kaggle competition?

6. 1 Learn from the best.
Forums
Scripts
Solutions from prize winners

7. 2 Work with cool datasets.
Flights in GE Flights Quest
Driver telematic analysis
Amazon employee access
rights

8. +1 You can also win money

9. What kinds of
competitions Kaggle has?

10. Public
In-class
Private

11. What languages can you
use?
Any open-source language
(sometimes also sponsor’s
proprietary languages)
Gnu Octave
(no Matlab)

12. What is the data like?

13. Data comes from companies
and non-profit organizations

14. Data sizes vary
Zip ~1 MB
Zip ~6 GB

15. Data comes in all shapes
Customer data
Log files
Timeseries
HTML pages
Images
Documents

16. How does Kaggle
competition work?

17. Competition flow
Duration typically 4 to 8 weeks
Max. 5 entries per day

18. test
Model
train
submission.csvPredict
Build prediction model
Calculate CV to
cross-validate

19. Evaluate submission
Typical evaluations
Area under the ROC curve
Normalized Gini coefficient
RMSLE
…

20. Public leaderboard
Private leaderboard
~10-30% of test data
submission.csv
Submit entry

21. Choose two entries for final
Practical choice
Best entry in public leaderboard
+
Best CV from local entries

22. Tips

23. Look at data. Visualize it.
Source
https://www.kaggle.com/justfor/liberty-mutual-group-property-inspection-prediction/explore-data/notebook
https://www.kaggle.com/odiseo1982/liberty-mutual-group-property-inspection-prediction/compare-variables-between-train-and-test/files

24. Focus on feature engineering
Feature selection
Feature construction
Dates
Locations
Categories
Segmentation
Statistics

25. Build different models
3 target variables
4 cities
=
Build 12 models
Source
https://www.kaggle.com/c/see-click-predict-fix/visualization/1390

26. Try different algorithms
Random forest
Vowpal Wabbit
GBM
Xgboost

27. Build ensembles
Average of submissions
Weighted average of submissions
Ranked average of submissions
Stacked generalization
Blending

28. Keep track of your
submissions
Submission id

29. Next steps

30. Start competing
Create Kaggle account
Choose competition
Go for it!

31. Useful links
Kaggle Blog
http://blog.kaggle.com
Kaggle Competitions: Where to begin
http://www.analyticsvidhya.com/blog/2015/06/start-journey-kaggle/
Kaggle Feature Engineering
http://machinelearningmastery.com/discover-feature-engineering-how-to-
engineer-features-and-how-to-get-good-at-it/
Kaggle Ensembling Guide
http://mlwave.com/kaggle-ensembling-guide/