Beginning your Kaggle Journey
Make submission to your first Kaggle competition
Kaggle, the popular platform for data science competitions is something that every data science and machine learning enthusiast comes across on their journey. Kaggle offers a myriad of datasets, hosts competitions with monetary rewards, provides courses for beginners and offers a no-setup Jupyter notebook environment with free GPUs.
Getting started with Kaggle is a solid idea as it allows you test your skills on real problems. Not only that it’s a great place to observe and learn from other experienced data science and machine learning practitioners and improve yourself.
However, entering Kaggle competitions for the first time can be scary. When beginners see teams of Ph.D. researchers and machine learning professionals competing, the feeling of inadequacy starts to creep in. I know this because I, too, am a beginner who felt inept as I browsed through the many competitions and submissions on Kaggle.
But the point isn’t to get top rank in your first Kaggle competition but to improve yourselves by learning from others and getting out of your comfort zone. I recently made my first submission and decided to write my first blog to help others do the same.
Getting Started
The very first thing to do is creating a Kaggle account if you don’t already have one. You can access the Competitions tab on the left after logging in. Here you can find all the ongoing competitions. You can also search and filter the competitions to find the one best suited for you.
It’s a good idea to start with an easy competition to get comfortable with the platform. On the competition page you can read through it’s overview, take a look at the data and also explore other people’s code. Once you’re a comfortable with the competition and understand it’s goals and evaluation metrics you can click on the Join Competition button to enter the competition.
Now you can either
- Download the data from Data tab
- Work with it in your own environment
- Make the final submission in the specified format
OR you can use Kaggle’s integrated no-setup Jupyter notebook environment to make a submission.
To create a notebook simply go to the Code tab and click on New Notebook button. This will create a new Jupyter notebook with most of the necessary packages already installed. So you can focus on just writing the code.
Running the first cell will show you the path of all the files required to complete the competition. You can now freely explore the data and experiment with various machine learning algorithms in order to meet the competition’s objectives.
Making a Submission
The only thing left to do now is to submit your work and check your results. But first, make sure that you’ve exported your output in the correct format, as stated by the competition.
Click on the Save Version button. Make sure the Save & Run all (Commit) checkbox is checked, then click Save.
Once the task is finished running, click on the version number next to the Save Version button. This will display a preview of the current version as well as a history of all prior versions.
You can now either click Go to Viewer or, if you want to open some other version, click the three dots and then select Open in Viewer.
To submit your work, select Submit to Competition from the three dots in the upper right corner. Make sure the correct output file is selected before clicking on Submit.
All Done
Congratulations!! 🎉🎉 This is all it takes to submit your work to a Kaggle competition. Now you can view your public score and ranking by going to the My Submissions tab.
It wasn’t so difficult, was it?
Kaggle competitions are great because you will always win, if you look at them as opportunities to learn and improve your skills. You can examine the top submissions and try to figure out what they did and why they did it.
Finally, don’t be scared of low ranks and get your hands dirty. After your first submission, you’ll be at ease with the platform and well on your way to obtaining the coveted Grandmaster title.
“The secret of getting ahead is getting started.”