My Kaggle Journey from Novice to Notebooks Masters
Ever got overwhelmed by Kaggle, or want to know about ways to succeed in Kaggle? Don’t worry, this blog covers it all.
Kaggle constitutes an important part of every data scientist’s career, being active on Kaggle not only helps you to build a great career, but most importantly it’s a place where one gets to learn from the best in the field and build connections for life. For those of you who are not aware of Kaggle, don’t worry I will be explaining the functioning of the platform in a bit.
I have been quite active on Kaggle for the past 6 months, learning a lot from other’s work and showcasing my work to the outside world. There are five tiers in Kaggle, namely Novice, Contributor, Expert, Master, and GrandMaster and there are various goals that you need to fulfill before you move from one level to another. I became Kaggle Notebooks and Discussion Expert on 27th January 2021, and to be honest, I thought that was the maximum tier that I could ever achieve, but after working hard for some more time, I finally became Kaggle Master on 25th June 2021. The feelings that I had on that day can’t be expressed in words. My current rank is 235 out of 176,033 in the world.
The main aim of this article is to share my Kaggle journey with you, how I overcame my fears or imposter syndrome, how I learned new things from the platform, and to encourage you to start your own Kaggle journey. Before we discuss these things, let’s discuss in brief about Kaggle.
What is Kaggle?
Kaggle is a community that allows all the data science enthusiasts around the globe to gather and take part in various competitions, upload datasets for others to experiment on, share notebooks or kernels and participate in discussions about Machine Learning, Deep learning, or anything in the field of Artificial Intelligence.
There are four different categories in Kaggle:
Competitions
You can team up or participate individually in different competitions posted by various companies or Kaggle itself. Competitions allow you to enhance your overall data science skills because you will have to do feature engineering, feature selection, and hyperparameters tuning, etc to select the best model and make final submissions to the competition. Competitions allow you to discover your true potential and make you learn many new approaches to deal with a problem.
Datasets
If anyone finds any unique or interesting dataset, he/she can upload it on Kaggle and can also mention some tasks like performing Exploratory Data Analysis and finding out answers to some questions related to the data uploaded, etc. Many times, if you have to perform some Machine Learning tasks for which no dataset is available, you can always refer to the datasets available on Kaggle.
Notebooks
Once you find out any interesting dataset, you can make notebooks or kernels for the dataset. Notebooks are highly interactive multi-purpose tools that allow you to write and execute code. You can explore the notebooks of other people as well, and for me, this is the best part of Kaggle, because when I am out of ideas or have some doubts about the approaches, I can learn from other’s notebooks and come up with new ideas of my own.
Discussions
In the field of Data Science, often you can come up with many doubts related to certain concepts or any execution-related queries, for such doubts discussion forum is the place to be. You can post your doubts, learn or answer other people’s doubts that you might not have. People will notice you as you ask more doubts or resolve doubts with time.
Whenever you do some activity in any of the four categories, other people can upvote(or like) your work. You get medals whenever your work crosses a certain number of Upvotes. There are three different kinds of medals: Bronze, Silver, and Gold.
For example, if you have made a new notebook, other people can check out your notebook and upvote it if they found it useful. For notebooks, the condition to get a bronze medal is to have 5 upvotes, 20 upvotes for silver, and 50 upvotes for gold, and this number varies for each category. As you get more medals, you advance from one tier to another. You can refer to the progression system here.
Whenever you sign up for Kaggle, you are considered a Novice for all four categories. There are some tasks and information that you need to complete to become a contributor. After that, the actual fun begins. You get a rank in respective categories only when you become an expert. Let’s say you want to achieve the master tier for Notebooks, first, you would need at least 5 medals to become an expert(at this stage you will get a rank for notebooks category), after that if you gain 10 silver medals in total, you will become Notebooks Master! More the upvotes you get better will be your rank. Likewise, there are different criteria for each category.
It’s okay to Get Overwhelmed
All of these things might look too overwhelming, and honestly, I could not start my Kaggle journey after reading about Kaggle. I wondered if I am good enough for the platform, whether people will like my work, and what not!
But believe me, it’s only about making that first notebook, asking or maybe solving that one doubt in the discussions forum. After that you will soon realize how welcoming the people are on Kaggle, you will discover new approaches to dealing with a dataset or maybe learn a new ML algorithm! The possibilities are endless.
Points to Succeed
Following are some of the elements that you can keep in your mind during your Kaggle journey:
Don’t just make anything for the sake of getting medals
A lot of people just make notebooks for the sake of getting medals, without taking care of the quality of their work. Please remember that whatever notebooks you make on Kaggle can be used to showcase your skills during a technical interview, if your quality is excellent and showcases your true potential, the employer might get impressed as well. Plus better the quality of the work, chances of getting upvotes increase.
Be Patient
A lot of times when you make your notebooks public, you might not get upvotes immediately, this does not mean your work isn’t good so don’t lose confidence. You will ultimately get upvotes at the end if your work is good. You can also promote your notebooks on Social Media Platforms like Linkedin, Twitter so that more people can check out your work.
Bookmark the interesting and amazing works
Whenever you come across an amazing notebook or an interesting discussion, do bookmark them to use that concept in your future work(don’t forget to give credits to the original creator). For example, I came across a code snippet that embedded photos of football players in the plotly bar chart for visualization purposes, I had never imagined that was also possible, so I immediately bookmarked it and will use it whenever an opportunity presents itself.
Interact as much as you can
Whenever you are reading some notebook or participating in any competition, do indulge yourself in interactions with other Kagglers, there are several benefits to this: first, people might check out your profile if you show appreciation towards their work, second, you will be well known in the community(this does not happen in days). Whenever you check out a person’s Kaggle profile, make sure to connect with them on Linkedin or Github, or Twitter, as they might help you in your time of need.
Explain your work
It is always good to write out comments for each step that you are performing or the reason for performing that step. Let’s say that you have plotted a scatterplot for visualization purposes, you can write out the inference from that plot so that other people can also understand your thinking process. Writing good notebooks is no less than portraying a story.
Succeeding in Kaggle is not one day’s job and there is no end to it, it is dependent on how you want to make the maximum benefit out of it. One piece of advice that I would give is to set up short-term targets, for example when I started, my first target was to become a discussions expert, once that target was achieved, I wanted to become a notebooks expert. After becoming an expert, I set the target of becoming Kaggle Notebooks Master. My next target is to participate in competitions. Therefore, I have always benefited by setting short-term goals and never got overwhelmed or got scared due to Kaggle.
Feel free to connect with me on Linkedin, Github, or Twitter.
Hope you can get inspired by this article and get started with your Kaggle journey. Here are some of the datasets or competitions that you can try to get your hands dirty with:
