How to Make Effective Kaggle Profile and Use as Data Analyst Portfolio
Hey! Those who read our posts are often well aware that we have devoted a huge amount of material to Kaggle. We talked about how to promote your profile (link), why this is the best platform for gaining knowledge and skills in Data skills (link), and also criticized developers for some of the platform's shortcomings (link).
But today we would like to supplement the collection of Kaggle material on our website and talk about how to properly design your profile in such a way that it was a full-fledged portfolio for applying for a job. I will talk about the best practices that I use myself, as well as those that I spied on from other users.
Duplicate your Code
As you know, Kaggle has the ability to attach links to GitHub and Linkedin to the profile. Be sure to do this if you haven't already, otherwise you won't be able to be contacted. After linking the profile to GitHub, duplicate all the code you wrote there. This will allow the employer to look at your code, send it to someone for evaluation, or simply download it to run the script on their device.
How I Dublicate my Code - Kaggle
This is especially true for data parsers. If you have written a script that collects data in a separate dataset, then you must publish the script on GitHub. When I parsed data from sites, I initially published the script code on Kaggle itself, but this is completely inconvenient. It is inconvenient to publish code on Kaggle that refers to some third-party resources. Yes, it is readable, it is understandable, however, it will not work for a potential employer to run the script and see how it works.
Moreover, duplicate code snippets. My profile on GitHub has a lot of Python code just for visualizing data, for decorating graphs, charts and other things, simply because I forget this syntax very often, and I need a resource where I could peep the syntax - GitHub is great for this.
Notebook on Kaggle also needs to be formatted correctly. Do not forget that, firstly, we make a portfolio that a potential employer will see, a person who will firstly pay attention to the appearance of the code - how it is designed, how many comments there are in it, How readable is the code?
How I Design Table of Contents - Kaggle
Be sure to make headings in notebooks. So, while reading, you will be able to view your code by the table of contents in the sidebar. Be sure to add comments to the key points of the code, explain what you plan to do next, explain why you are introducing this or that variable, explain why this or that graph is needed, what it shows what you want to do with it.
The fact is that answering the questions above is a huge part of a data analyst's job. When analysts compile dashboards for datasets, they are required to answer approximately the same questions for themselves. That is why it is very important for a potential employer to show an understanding of what you are doing, and most importantly why, because this is your part of the work in the future.
Take Part in the Competition
Be sure to take part in the competition held by Kaggle. It is not necessary to take high places or win at all. You can take at least the last place - it does not matter. It is important to simply show that you want to try something new, that you want to join this community and show yourself. Be sure to find yourself a team, at least one more person who would share your interest with you. How to do this, we have a separate article.
Kaggle Competitions Page
Interaction with another Kaggle user, an expert, will show you as a person who is sociable and ready for team decisions. These are soft skills that people so often ask about in interviews. In these competitions, victory is not important, but how you, for example, divided responsibilities, how you communicated and interacted, how seriously you approached the competition. All this can be written in the notebook itself.
Very few people prescribe such things, which is very in vain, because other users, like employers, are interested in how you divided the responsibilities and which of you was responsible for which part of the study. This is the experience of teamwork, which can be safely indicated in the resume, this is the experience of project work, which can also be indicated in the resume. As you can see, there are only pluses around, even without getting a high place in the competition.
Don't Waste Time
Do not waste your precious time promoting your account in third-party resources such as social networks. For example, I used to think that the more visibility of my profile, the more likely that someone will notice my profile on Kaggle, but this is not the case at all. The fact is that Kaggle is designed in such a way that the dataset or user code will be looked at and noticed by everyone who needs it.
You don't need any promotion, you don't need any social media posts. All you need is to be active on Kaggle. Help other members in the discussions section, comment on whose code or dataset you like. In a word, in order for your profile to be more noticeable among others, you just need to be active. Everything is simple.
No hot topic datasets will help you here. While ChatGPT was gaining popularity, a dataset about 2022 wheat harvests was trending on Kaggle. I say this because many people are trying to publish a very up-to-date dataset in a certain period of time, get a lot of upvotes and comments, thereby getting a local medal. But most of the time it doesn't work. Post what you personally like, then your profile will become more popular.
Most users use the discussions tab solely to ask questions to other users of the platform, or to answer and help others. That's great, but this tab can be used as a place to post personal tutorials and small articles. It is not necessary to publish scientific dissertations, it is enough, for example, to publish pieces of code and a small author's explanation on this piece.
For you, this may seem useless, because you wrote this code, but for others it is not, for others it will help solve a specific problem that they encountered while writing code, visualizing data, or building predictive values.
I find that writing such mini tutorials is much more effective than answering user questions that are the same. In addition, such tutorials do not need any SEO optimization on your part and appear on the first pages of the search engine, which also adds attention to your profile.
I would really like you to use your Kaggle profile as your portfolio, because everything you need is there. On Kaggle there is an opportunity to show yourself, show what you can and love to do. Moreover, there is an opportunity for others to evaluate your work. If I were hiring a data scientist, I would pay attention to how much the candidate's code or dataset is liked by other members of the community. After all, this is also an important indicator.
Make your profile right, be active and don't forget that other people will see your code, so respect their time - write comments on your lines of code, make your notebook beautiful and bright, with pictures and memes, make sure that even your mom understands the code, which you wrote. Good luck!