Beginners Guide To Kaggle
Introduction
Kaggle is one of to largest data science communities for open-source data and collaboration. It attracts a huge number of data scientists with its competitions who want to earn the recognition of the data science community and be part of a good cause to use data wisely. Although Kaggle has many data science-related courses and beginners notebooks it can be quite challenging to navigate Kaggle as a beginner. This article familiarizes you with the basic things one needs to know to use Kaggle well.
Notebooks
Kaggle notebooks function the same way as Colab or Juptyer notebooks Kaggle has its own environment that has the most widely used packages preinstalled in the environment. Like Colab, Kaggle notebooks allow you to use GPU and TPU albeit for an allotted amount of time. At the time of writing this article, GPU time was 40 Hrs and TPU time was 9 Hrs. One of the best features about Kaggle notebooks is that it allows you to use datasets stored in Kaggle saving one the hassle of Gathering and Storing the Data to the cloud. Add the data or models that you want to use in the notebook.
Datasets
Kaggle Datasets have convenient features that are not available on other online Notebooks such as Jupyter or Colab. Datasets allow you to upload your data online and makes it easier to share with other data scientists. The outputs from notebooks can also be used to make Datasets or update existing ones.
Courses
Kaggle offers a variety of courses aimed at beginners. From python to Neural Networks and Reinforcement learning these courses are a great way for beginners to learn the basics of Data Science but get less useful as you progress in the field. The courses are a great way to get started on Kaggle.
Discussions
Kaggle boasts of a wide user engagement allowing you to connect yourself to a wide range of users. You can also comment on other notebooks offering your opinion and critique on their code. Discussions are a good way to build contacts in the field and answer your queries.
Competitions
Competitions are one of the main focuses of Kaggle and are the primary reason why both Professionals and Beginners join Kaggle. It offers Cash prizes that bring together the brightest minds to solve challenges in data science. When participating in a competition you will be exposed to the most advanced techniques used to solve a problem. While most of these solutions are fantastic there has been controversy regarding the usefulness of the solutions to the hosts of the competition. Since the competition requires the user to obtain the highest score almost all competitors focus on boosting their score ignoring the fact that this solution will have to be used on a practical basis.
Medals & Rankings
Users can upvote the Datasets, Notebooks and Comments(Discussions) that they like. Receiving a certain number of upvotes or coming within a certain ranking in a competition gives you a bronze, silver or gold medal. As you progress in Kaggle engaging with the community and winning medals you will increase your Kaggle tier from Novice -> Contributor ->Expert -> Master -> Grandmaster. Grandmasters are given special privileges and are allowed to take part in exclusive competitions.
Conclusion
As you can see Kaggle is a great place for both Beginners as well as Veterans in the field to engage. If you are a beginner I would recommend starting with the courses and then moving on to public datasets. Lay off competitions until you feel comfortable. Competitions tend to use advanced concepts which might be jarring to beginners. Try solving problems that you feel comfortable with such as datasets and discussions. This is a great way to build up your knowledge and build contacts in the field.
