Develop a Stand-out Data Science Portfolio with GitHub

Image generated with ideogram.ai

In today’s data science work climate, there is so much competition for even the entry-level data scientist position. There is a surplus of applicants, yet few positions can be filled. The imbalance in supply and demand in the data science field causes many freshers or career-switchers to give up on pursuing data science.

So, it’s pretty much up to the applicant to stand out from the others. However, how can we do that? Your portfolio is your way to stand out.

With so many people aiming for the same position, the portfolio you develop can be a differentiator that allows you to get the job interview.

In this article, we will discuss the tips on developing a stand-out data science portfolio with GitHub.

Let’s get into it!
 

Data Science Portfolio with GitHub

 
GitHub is a renowned cloud-based platform that stores and shares code with the community. It uses the repository concept to showcase or share our works. The platform also allows the users to track the changes in the repository over time.

Modern data science uses a programming language to perform our work. Using coding language means we could use GitHub to share our works as the portfolio.

It’s easy to upload our data science code to the GitHub repository, but it will take a lot of work to make it stand out.

To develop a stand-out data science portfolio with GitHub, we will need to focus on showcasing a combination of our technical skills, problem-solving abilities, and creativity. Let’s see each tips that you could have to improve your portfolio.

Setting Up Your GitHub Profile

First impressions always matter. It’s just like your CV; the first time someone looks at your GitHub profile, they should have the impression that you are professional and ready for data science work.

Develop your GitHub profile that shows your experience, background, and interests. You can also add a profile README as a dedicated page to summarize your skills and project highlights so the potential employer can understand you better.

For example,Sajal Sharma’s GitHub profile showcases simple and quick information about his professional background as a data scientist.
 

 
His profile shows his experience, education, tech stacks, and everything necessary to the employer.

You can also pin the projects that you want the people to see.
 

 
By pinning the project that you find essential, you could help potential employers understand which project they should assess to learn more about you.
 

Selecting Project to Showcase

I am sure many applicants want to showcase all the projects you already do as you are already working hard for them. However, we still need to select the projects that need to be presented at the forefront to represent you as a data scientist.

First, you should present only projects that showcase your ability to solve real-world problems. Avoid a sandbox project such as Iris or Titanic, as it’s already synonymous with a practice project rather than something that fully shows your skill.

Second, focus on a few high-quality projects that have been finished rather than numerous unfinished ones. Also, the projects should show diverse skills you have, such as how to perform machine learning development, data cleaning, visualization, etc.

Lastly, select a project that showcases your knowledge of the business domain you want to enter. This could be finance, technology, healthcare, or any other domain. Showing a project relevant to the company you wish to apply to will undoubtedly improve your chances of standing out.

For example, James Green’s GitHub portfolio shows a simple visualization and summary points of the results of his stock portfolio selection tool project.
 

 
You don’t need to explain all the details in the GitHub repository. Still, you will at least provide a link to the project documentation or article to explain the project more precisely.

Another example is the Francisco J. Carrillo’s Project Portfolio, where he showcases his Prediction of Stochastic Clogging Processes project in a simple yet meaningful manner. He describes what happens in his project and the required skills while still attaching the notebook link to learn more about his project.
 

 
Try to select the project that will represent you the best and could showcase your skillfully.
 

Structure Your Project Properly

One essential part of making your Data Science Portfolio stand out is having a proper structure on GitHub.

I mentioned above that you can link your detailed works in another form, such as a notebook and blog, but you can also use your GitHub repository to explain your project.

With the repository README for each project, you can explain your project’s goal, the result, the technique used, and many other relevant things to showcase your data science project in the repository. To make it stand out, we need to structure how data scientists will work better and make it easier for people to follow your project.

Let’s take an example of the Shanthibooshan Subramanian portfolio, which structured his project with a simple table of contents.
 

 
Additionally, Arch Desai’s project portfolio shows a great structure. He prioritized project organization so readers could better understand the portfolio.
 

 

You can use the Cookiecutter Data Science Template to better structure your data science project.
 

Conclusion

 
Developing a stand-out data science portfolio with GitHub is necessary to differentiate ourselves from the other applicants.
In this article, we have discussed several aspects of improving your data science portfolio, including setting up your GitHub profile for first impressions, selecting relevant projects, and structuring the portfolio better.

I hope this has helped! Please share any tips you feel are necessary to make your data science portfolio stand out.
 
 

Cornellius Yudha Wijaya is a data science assistant manager and data writer. While working full-time at Allianz Indonesia, he loves to share Python and data tips via social media and writing media. Cornellius writes on a variety of AI and machine learning topics.