5 Tips on Providing Valuable Feedback on a Data Science Project

A lot of us don’t feel comfortable giving someone else feedback on their work. We believe that we either don’t know enough or aren’t an expert to help someone else out.

But you don’t need to be an expert at all! In fact, attempting to give feedback on someone else’s project also helps you learn and grow more on your journey.

So, here are 5 tips on providing feedback in the community -

Compliment and Encourage

Always aim to compliment someone on their work. It’s a difficult journey to complete a project, but it’s even scarier to share it with others and ask for feedback.

Praise someone for what they have done well and encourage them to keep progressing with the same tenacity.

Here’s a great example from Nityesh, Community Manager at DataQuest, on a student’s guided project -

  • He praises the project.
  • Shares a couple of aspects he liked.
  • And points out student’s existing achievements as means of encouragement to share more of their work.

This not only helps them, but it also reminds you that your own projects are worth praising even if you are a beginner.

Constructive and Actionable Feedback

It’s always great to know when you did well, but all of us want to go further. To improve ourselves and learn more.

You can help someone else by being constructive and actionable with your feedback. Providing actionable comments can be difficult, especially for beginners. But spending a bit more time focusing on asking yourself questions about their project can help you narrow it down.

From the same guided project as above, a fellow student shared some valuable feedback -

  • The feedback points out one aspect of the project to focus on, and
  • The suggestion on keeping the same plot orientations is actionable.

This will also help you provide feedback to yourself on your own projects down the line.

Genuine Curiosity

Working in any data-related job means you look at the data and ask questions about it. There is no better way to understand and dive deep into understanding the problem you are trying to solve than asking questions.

Many of us, get too focused on just completing the project or maybe even just following the given instructions. So much so, that we fail to stop, think and reflect on our work.

When providing feedback to other students, try to come up with questions that make the student think -

Oh, wow. I hadn’t even thought of that.

Asking such questions can be difficult. Here are some standard approaches you can take -

  • “You did X, what if you did Y instead? What do you think would happen?”

    • For example, “You calculated the mean of those prices. What if you calculated the median instead. How do you expect your observation to change?”
  • “You gathered insight C. Why do you think that happens?”

    • For example, “You found out that Free Users spend more time on the app than Customers. Why does that happen? Shouldn’t it make more sense that Customers utilize the app to their fullest? Is there anything else in the data that could help answer that?”

Such questions aren’t easy to come up with. The best way to do so is to have worked on the project yourself first to understand the data and ask yourself such questions as well. Be genuinely curious about the data and share your curiosity with others through those questions. Start with the “whys”.

Here is another great example from the community -

  • A specific question that can allow the student to extend their analysis.

This is especially helpful because you will one day be collaborating with team members on a project. Being able to brainstorm through such questions will be a frequent occurrence.

Share Valuable Guides

We should never expect a highly refined and polished project from any learner. However, each and every one of us should aim to iterate over our projects. They have far more impact when we later add them to our portfolio.

Luckily, DataQuest offers an excellent Style Guide that you can share with students on how they could improve their project.

But it’s important to note that it’s not about just sharing the Guide. Try to use a particular point from the Guide and compare it to the student’s project. This relates back to sharing actionable feedback. Help them understand how they could improve one part of their project, and then they will have the freedom to iterate on it later.

Follow Up

All of us are busy; all of us want to keep progressing in our learning journeys. But all of us also feel confident in ourselves when we know people are invested in our journey as well.

Try to follow up with the student on their project after you have provided some suggestions to them. After a few days or a week or two. Ask them what iterations they were able to make to their project and encourage them to share it again if they want some more feedback.

Here is a response from a student when Elena, Community Moderator, decided to follow up -

You can see the impact her follow up had on the student. I would also recommend checking out that entire thread. Elena’s feedback is a gold mine!


There are different ways to provide feedback and the above are definitely not exhaustive. But each step is meant to help both the student and yourself through your learning journeys.

I hope some of you might be able to use the above guide in meaningful ways! What would you add to the above?


I write about online education, learning, and exploring my curiosities. If you’d like, you can connect with me on Twitter.


Photo by Charles Deluvio on Unsplash

7 Likes

Wow, I absolutely love this @the_doctor! :heart_eyes:

Coming from one of our most active Community Moderators, these tips are gold.

Thanks for sharing them with the community! :heart:

1 Like

Hello @the_doctor! This is a great article. Even though, I have some experience in project reviewing I now see some points I can improve like trying to ask the student more questions about the data they worked with or pointing out a specific part of the Style Guide!

I would add another point: ask the student about the ways they think that can improve their project:) We can also collect a series of excellent GPs and share them as examples to learn on… But I’m not sure if it will be helpful. What do you think, @the_doctor?

1 Like

Thank you @the_doctor for sharing such a cool and useful article with the Community! :heavy_heart_exclamation: Agree with each point, everything totally makes sense! I would also add the following technical points to which reviewers can pay attention while reading other people’s projects:

  • An overall project structure: title, subheadings, introduction, conclusion, coherence of numbering, if present.
  • Storytelling and observations in markdown.
  • Keeping the code clean, maybe suggesting some more elegant code solutions for some cases.
  • Efficient code documentation (code comments and function description).
  • Dataviz best practices: everything that improves graph readability (including titles, labels, fontsize, annotations, color selection, despining, de-ticking), increases data-ink ratio and minimizes chartjunk. I remember it was @artur.sannikov96 who suggested me to despine the graphs in one of my first projects here, and now I myself advise it to everyone! :smiley:
  • Various issues that I don’t know how to call collectively :slightly_smiling_face: For example: typos, not rendered images for some reasons, commented-out code left in the project, forgetting to re-run the whole project, links that don’t open, etc.

Also, thank you for your appreciation of my project reviewing activity! And you’re absolutely right: it really helps to grow quickly ourselves and to learn a lot of things, at least in my case it happened exactly so!

2 Likes