How to Present a Data Science Project

How to Present a Data Science Project


After passing a company’s take-home challenge, you might get asked to present your data science project to data scientists and the hiring manager. Presentations are high-pressure, especially if public speaking is not a strong skill for you.

Fortunately, making your data science presentation more engaging (and using it to land you the job) is a straightforward process. Whether you have a data science project presentation for a job interview or you are presenting the final project for a data science course, the key is to:

Data Science Presentations: Where to Start

Design your presentation for the audience and their goals. For example, if you’re presenting to non-technical stakeholders, your project shouldn’t be loaded with technical jargon. Or, conversely, if you’re presenting to a group of data professionals, don’t bore them with beginner definitions.

Before you put together a data science project, ask yourself these questions about the audience:

  • Who is your audience? How technical are they? Why are they attending the presentation?
  • What potential questions will they have about your project?
  • What types of data/analysis will be most interesting for the audience?
  • What do they want to learn about you or your work during the presentation?

For a Job: If this presentation occurs after a take-home challenge, usually you have 45 minutes to present, followed by 15 minutes of Q&A.

Don’t forget to prepare for the Q&A: List all of the possible questions the audience might have and develop answers for each of them.

What to Include in Data Science Presentations

Data Science Presentation

You should use the slides you create to tell a data story. At the very least, you’ll want to include the following in slides for your presentation:

  • Overview - Summarize the problem statement and convey the importance of the project. This could be split into two, with a brief overview of the problem on Slide 1 and a bulleted list on Slide 2 with the potential impact of your findings.
  • Methodology - Summarize how you approached the problem, including initial assumptions, clarifying questions you asked, challenges faced, and steps taken in investigating the problem. Keep this high-level, unless you’re talking to a non-technical audience.
  • Your Findings - Explain what you discovered. Did you find support for your hypothesis? How did your machine learning model perform? Support your findings with data, visualizations, key observations, etc. This is the most important information for your audience to have, so make this a focus of your presentation.
  • Recommendations - Answer these questions: What does your analysis say about the business? What recommendations would you make? Presentations are a chance to showcase how you would apply data science to the sample problem, and your recommendations show your product and business sense in action.
  • Conclusion - Reiterate important takeaways, but also take the time to discuss the next steps, such as if further analysis is needed, improvements you might make, or if you would have done something differently with more time or resources.

Designing Slides: Use clean, simple designs for your slides, including large headlines, very short texts (less than 20 words), and visualizations that help you tell a story.

Rehearsing Your Presentation: What to Do

Practicing provides a chance to work out any potential tech-related issues (slides, audio, and visuals) and speaking-related problems. During rehearsals, practice exactly what you want to say. However, keep it conversational.

Ideally, do some practice runs of the presentation for colleagues and record your initial takes. From there, work on refining the presentation and finish off with another session to polish your work.

Here are a few tips for getting the most out of your rehearsal time:

  • Create a script - Don’t create a word-for-word script. Instead, have speaking notes for each slide that provide a general idea of what you want to convey. Relying too much on a script will make your presentation sound over-rehearsed, and may trip you up if you end up deviating from it.

  • Do mock presentations - Present to friends and colleagues, and ask for feedback, questions, and overall comments. Ideally, you should practice with both technical and non-technical audiences. Their feedback will help improve flow, improve clarity, and remove extraneous info.

  • Record yourself - At a minimum, record audio of your practice, though adding video is even better. Review the audio for flaws in your speaking – Are you talking too fast? Do you say “um” too much? Video will help you review body language – Are you hunched over? Do you have your face glued to the slides?

  • Rehearse the Q&A - Forgoing this step is a big mistake. It’s not a good look to nail the presentation, only to bomb the Q&A right after. In particular, you’ll want to prepare answers to questions about your models like:

    1. Did you have any benchmark performance to compare to?
    2. Why did you choose the model you did? What were the limitations?
    3. Were there biases? How did you account for these biases?
    4. How would you improve the model?
  • Do a tech run-through - Practice using your slides, audio, and video. If this is a video-based presentation, do all of your mock presentations via video conference. Practice using a microphone or headset, ensure you are properly lit, and practice sharing your screen.

Tips for a Data Science Project Presentation

As you build your presentation slides and rehearse, here are some of the best practices and tips to make your performance even stronger:

  • Keep it concise - Keep your presentation simple and to the point. You can’t show every step you took. Instead, keep it brief and to the point, focusing only on key details.

  • Choose your best visualizations - Images and charts make your presentation easier to follow and clearly display the impact/findings of your project. Include only vital information in the chart, and be sure to consider fonts, color theory, and other good practices of visualization design. A general rule of thumb: It should be clear to a layman what a chart is conveying.

  • Focus on the impact - If you’re presenting on a project from a previous job, show the impact it had using metrics. Increased revenue, reduced churn, customer acquisition, and other factors will illustrate how your work impacted the bottom line.

  • Include limitations - Every project has limitations and challenges. Although it might seem counterintuitive to talk about what went wrong, discussing limitations will make your presentation stronger. It shows you can identify potential flaws in reasoning and that you care about quality controls.

  • Talk through your decisions - Explain why you made the technical decisions you did. This will help the audience understand your approach, what factors lead to you making a certain decision, and how you personally use creative problem-solving.

  • Make it accessible - Explain the technical details of your project in layman’s terms. Examples and analogies can be helpful for audiences, and ideally, you should be able to explain an algorithm or complex data science technique in one or two sentences for a non-technical audience.

For the Presentation: Final Tips

Public speaking is nerve-wracking. But there are strategies you can take to calm your nerves and make the most of your presentation time. Here are public speaking tips for your data science presentation:

  • Make eye contact - Eye contact connects you with your audience and makes your presentation more engaging and impactful. One strategy: sustain eye contact with one person per thought. Be sure to practice this during your rehearsals.

  • Allow space for questions - Although there’s usually a Q&A at the end, questions can come up throughout. If you’re not sure if the audience has questions, take a pause and ask, “Does anyone have any questions?” Remember, you don’t want to talk AT them.

  • Avoid rushing - Focus on pacing. You should be talking at a normal conversational speed. Too fast, and you’ll end up losing the audience. Too slow, and you will bore them.

  • Breath, relax, and collect your thoughts - Before you begin, take some deep breaths. One strategy: reframe the focus from you (e.g., “What if I blow it?”) to the audience (“My focus is helping the audience understand and learn.”).

More Data Science Interview Prep Resources

If you’re looking for data science project ideas, see our guides for analytics projects and machine learning projects. You can also practice for your interview with these resources from Interview Query: