So you have started your machine learning/data science course. But still, you are not confident enough on how to solve a real-world problem or maybe you want to go deep in this field but don’t know where to start or how to get datasets, where to build the mindset like a data scientist, or how to show your skill to the rest of the world, or maybe you just want to practice on different datasets and expand your knowledge, then my friend Kaggle is the right place for you to get started.
What is Kaggle?
Kaggle is an online community of data scientists and machine learning practitioners. It is a platform where users find and publish their datasets, they explore and build a machine learning model in a web-based data-science environment. Kaggle provides a medium to work with other data scientists and machine learning experts.
Kaggle is most famous for its competition where companies upload their problem along with dataset and competitor around the globe solve their problem using AI/Machine Learning. And the winner of the competition wins the prize. Kaggle has over $1,000,000 prize pools. Here is the screenshot of the competition list and money which they offer on winning.
It consists of more than 19,000 public datasets and over 200,000 public notebooks. People who compete on Kaggle are often called Kaggler.
Kaggle provides many services let’s look at them one by one:
Machine Learning competition
This is Kaggle’s first and most famous product for which kaggle is known for. Here Companies put problem and machine learner/data scientists fight against each other for the Best Algorithm. The winner of this competition gets cash offered by the Company.
There is a number of competitions offered by Kaggle:
These are the competition for which Kaggle is best known for. Big Companies, Organizations, Government sponsors this kind of competition. They offer cash going as high as a million dollars.
These kinds of competition offer problems which are more experimental than competitive problems. They may or may not offer money or points due to their experimental nature.
These are more starter friendly competition or to put it in layman term these competition are for newbies who have just started practicing Machine Learning. They contain a simple dataset and have no deadline. Also, they don’t offer any prizes or money.
We can say that these competitions are of intermediate level. They may offer small prizes. One simple example of this competition is Digit Recognizer.
One of my favorite feature of Kaggle is it provides inbuilt Kernel. They nothing just Jupyter notebook in the browser. These Kernels are entirely free, you can also use their GPU to train large dataset. This saves you the hassle of setting up a local environment and also if you have a low configuration system where training your datasets takes longer you can use these Kernel to train your dataset without buying a new system. Kaggle kernels support many different languages but most popular are Python and R. Kaggle Kernels are publicly available to everyone so you can also read kernels of other people. This is a great way of learning new techniques and also getting involved with communities.
Public Dataset Platform
Kaggle is the best platform to find, discover, analyze open datasets. You can find many interesting datasets of a different type, different sizes from which you can improve your machine learning skills. You could even upload your own dataset. It has a dataset of everything from bone x-rays to results from boxing bouts.
It also provides free micro-courses. These courses are such that they train you to apply your domain knowledge to practical data. They are the fasted way to become data scientists and improve your skills. This is the fastest way to become a data scientist and improve your skills. Some of the micro-courses provided by Kaggle are: Python, Intermediate Machine Learning, Data Visualization, Deep Learning, etc
These are where you ask a question and get answers or solutions from thousand of the data scientist in the Kaggle community. There is typically six general Discussion form :
- Kaggle Form This form is specific for Kaggle community. This form consists of upcoming events, topics, etc.
- Getting Started As the name suggest this form is for people who are new to Kaggle.
- Product Feedback This is a feedback forum. Here people will generally give feedback to improve Kaggle community.
- Question And Answer These are where you will get answers/solution from other machine learners and data scientist
- Dataset Here you can request you on type of datasets.
This is also the best place to discover machine learning/data scientist jobs. Here employers post machine learning and AI-related jobs.
So what are you waiting for, sign up for Kaggle and improve your machine learning skills?