topic: Portfolio of work - tech guidelines

Tags employability-sprint github

GitHub Profile Best Practices

To assess the quality of your work, the hiring manager needs to access your raw code, which makes a solid GitHub profile essential. It needs to show the meaningful and impressive work you’ve done. The code should work reliably, and it needs to be clean and readable.

This article covers some of the basic requirements for your GitHub profile, as well as some really great advice on keeping your profile active: 5 tips to power up your github profile

You don’t need to go overboard with the aesthetics, but a little can go a long way – particularly with the first-line recruiters who might be scouring through hundreds of profiles. This article covers some easy ways in which to add a little polish to your GitHub profile: How to create a stunning Github profile

Here is a curated collection of great GitHub profiles. Some are better than others, so take the ideas you like and lose the ones you don’t.

It’s easy to get carried away with too much fancy “polishing” of your profile, so we highly recommend giving yourself a hard deadline on any beautification work. The most beautiful profile in the world won’t impress the hiring manager if the projects in the portfolio aren’t great.

Web Dev Portfolios

This article from the awesome team at freeCodeCamp outlines some of the minimum requirements for a Web Developer portfolio site. It then provides five great tips for powering up your profile beyond just the basics.

You can find some other great ideas and additional inspiration over here: 21 Best Developer Portfolio Examples

Data Science Portfolios

It is worth emphasising again: Your projects need to help you stand out!

Early-career data scientists are particularly prone to listing generic projects in their portfolios. Everyone has done the Titanic survival predictions or hand-writing recognition on the MNIST data set. These generic projects are great for your learning, given the huge community that’s built supporting resources around them, but they really shouldn’t be a core part of your portfolio. Nor should any other projects that come with a walk-through tutorial on YouTube.

Again, we suggest an iterative approach based on real-world problems. Find a topic that interests you and ask a question about it. Next, find data to answer your question. You can then publicly post your results and ask for some feedback. Rinse and repeat.

Even better – see if you can find a real client! Even if they don’t pay you just yet, it’ll be super useful to build something that adds tangible value in the real world and has meaningful deadlines attached to it.

Sometimes it’s better to learn in a team. Engage with some competition communities and see if you can find a team who’s willing to take you on. Ideally, you’ll want a team with at least a little competition experience.

Kaggle and Zindi are great places to start.

This article has some wonderful advice for data portfolio projects:

  • Use real data/scrape your own data/use publicly accessible APIs
  • Pick something you’re curious about, not something you hope will be impressive
  • Pick an analysis that is interesting, regardless of what you find
  • Perfect the visuals/keep the text short/make your data interactive
  • Productionise your analysis if possible/put the code on GitHub
  • Seek feedback (Remember the iterative approach to portfolio building: “Don’t be afraid to keep adding on to or editing your projects after they’re published!”)

This great article from dataquest.io provides a very useful overview of what employers are actually looking for in your portfolio. It also covers the basic types of projects you should include in your portfolio. (We’re not huge fans of the Explanatory Post – but otherwise, their advice is solid!)

Finally, for a bit of polish, we really like datascienceportfol.io as a quick and easy way to show off your projects in a concise, elegant, and visually appealing way. You might find some inspiration in the portfolios below:

  • FisherKK (GitHub) – a nice way to group similar projects, and to highlight coursework you’ve done.
  • https://otoro.net/ml/ – beautiful visuals! Plus a very clean and tidy structure, with wonderfully concise yet descriptive project overviews.

RAW CONTENT URL