What Makes A Successful Data Science or Engineering Capstone Project?

A capstone project is a major aspect of The Data Incubator, so crafting a quality project is critical, and in this article, we’re going to tell you what will make your idea stand out.

But first, what is the point? Why do a capstone project?

The main reason is to make sure that you experience all the individual parts of a data science project including ideation, planning, gathering the data, cleaning the data, doing analysis and then preparing a result that can be shared with other people.

Also, as a TDI fellow, you’ll be working through a series of mini-projects. These are where you prove your skill in a particular area of data science like machine learning or SQL.

The capstone is different because it is testing how you can pull together all of the individual elements and skills into a unified project. In the end, you should be able to show that you understand how these pieces relate and be able to use them to solve practical business problems.

The capstone project also serves a few other important purposes.

  • First, when you are interviewing, you can confidently respond to the question “how would you handle this particular problem as part of a data science project for our company?” Your project allows you to share your hands-on experience.
  • Second, the capstone is an important component of your portfolio. You can use it to demonstrate your technical skills. It provides prospective employers with a way to assess you as a candidate.
  • Third, you can include your capstone on Github. Not only is this the code you’ve written, but you can also show some of the end results as well.
    Finally, the highest quality capstone projects in a cohort will be a part of our capstone showcase. This happens every quarter where we invite our hiring partners to come in and see the best presentations.

What are we looking specifically for on our end?

The requirements aren’t specific because what makes a good capstone project varies and depends on the subject. It’s not as simple as a checklist. However, we have identified five elements that any good capstone project will have, and we’ll dive into each of these in this article.

But first, a quick look at some successful past projects.

Examples of Success

One student wanted to determine what the best ramen restaurants were in the bay area, and it turns out if you look at Yelp, all ramen restaurants have 3.5 stars. However, by looking at what people actually wrote about those restaurants the student was able to determine that there were a bunch of reviewers who had a variety of interests. And depending on your interests, you could recommend a particular restaurant.

Another fellow built a system that allowed users to find pets to adopt. They built a project that lets you upload a photograph, and they would find other similar pets that were available to be adopted.

Another project tried to predict flight delays, and they answered the question: “were you going to be stuck in an airport for longer than you expected?”

So these were all interesting and successful capstones, but how do you make that excellent one?

Element #1: The Project Must Solve a Real Problem

One of the things our hiring partners say again and again is they don’t want academic projects. The best projects have a real-world business application and they have an end-user.

As part of that, you want to consider who that user is that has this problem. This helps you craft your approach strategically, and it aids you through the decision-making process. When you’re faced with those 17 different things you could do at any given step, you’ll have that user persona to guide you.

One way to find that user is to be the user. So if you have a problem, then a user with that problem exists and you know precisely what that user is looking for in terms of a solution.

Some of the best projects we’ve seen are projects where people are trying to solve a problem that they had themselves. This makes it easier to get that personal connection. In general, it tends to produce more passion.

That excitement is contagious. If you’re excited about your project, the audience is going to become excited about it as well.

Element #2: Your Data Should Provide Novel Insight

We encourage you to find not just any problem but a unique problem to solve. Our hiring partners have seen the same sort of projects many times over, and you want to stand out a bit.

A couple of things you can do to be unique:

  • Find a new dataset. Find a data set that seems a little bit obscure or that not many other people analyzed. We have seen the New York City bike share data set used many times, it doesn’t mean that it’s a bad dataset, but it doesn’t stand out quite as much.
  • Use a common data set in a new way. The New York bike share data set is usually used to try to predict where bikes will be or where they’ve gone, but I’ve only seen one person use it to predict traffic past stores to understand the market size in those areas. Sometimes, a common dataset can be used in a new way that is exciting and innovative.

It helps to get started on this process earlier rather than later. Coming up with good ideas takes time, finding data that will actually help you answer those questions takes even more time. You shouldn’t want to wait until you get to this step on the application to start planning your idea and data.

We’re not expecting you to come in with a finished project, because if you did, you wouldn’t need to come to the data incubator, but we do like to see definite plans.

How are you going to go from this data that you’ve started gathering to a solution to that user? These plans will change as you gather more data as you see what it reveals, but that planning is essential.

Element #3: Turns the Project Idea into Working Code

Achieving this requires non-trivial data ingestion and preparation. Most data scientists estimate that they spend at least 80% of their working hours gathering, cleaning and preparing data for analysis. That’s not exactly the exciting stuff that everybody likes to talk about, but it is absolutely critical to your work.

There are three areas where at least one must be involved in any capstone project.

  • Machine Learning
  • Distributed Computing
  • Interactive Web Application

Many of our projects involve more than one of these areas. This is your opportunity to show hiring partners, your technical skills, highlight your abilities and show hiring partners why you’d be a great addition to their teams.

Element #4: You Include Several Types of Data Visualization

You need to be able to communicate your results. So, the capstone needs to have several different types of visualizations. You should take the time to make sure that these visualizations are clear and convey a story.

This will come into play later on as you are fine-tuning your project throughout the program. You’ll have assistance from our instructors to help you in this area.

Element #5: Your Work Should Yield a Useful Product

Finally, in the end, there has to be something you produce. Some deliverables could be a website, Jupyter notebook or a report.

We’ve had a variety of deliverables, and we don’t care too much about the exact form. In fact, you should think about the business objective and choose the form that best satisfies that objective.

What NOT To Do When Building Your Capstone Project

Knowing the right elements of a successful capstone project is helpful, but so is know what doesn’t make a successful project.

  1. Don’t choose a project based on the data set.
    Don’t try to choose a project solely based on a cool data set. Some of these can be successful, but in general, they tend to struggle to find the end-user. They end up going in a bunch of different directions as they’re exploring the data. If you are able, I would start with that problem, and then try to find the data sets that are going to be applicable to solving that problem.
  2. Don’t choose a project only based on the industry you’d like a job in.
    Don’t select a project based solely on the industry where you’re looking for the job. There are a couple of important reasons why you shouldn’t take this approach.First, keep your options open, consider a broad range of industries. You don’t know who’s hiring right now. Don’t restrict yourself too much.. You’re not looking for your dream job right now. You’re looking for something that’s going to give you good experience for the next three to five years.Second, hiring partners are interested in seeing how they can apply the things that you’ve come up with to try to solve the problems that they have. One student whose project focused on building a fantasy basketball team gained the interest of a hiring partner, not because they were trying to assemble their own fantasy basketball team, but they thought that the techniques that the student was using to solve that problem would apply to a problem that they had.
  3. Don’t choose a project based in order to use a particular technology.
    You’re being hired to solve problems, not to use a particular technology. It may seem very fancy to use TensorFlow to solve a problem, but if you could actually solve that problem with Excel instead it isn’t an impressive solution. Use the right tools to solve that particular problem.

How to Perfect the Capstone Pitch

You’ll need to pitch your presentation in the interview stage to earn a fellowship, but you’ll also have several opportunities to pitch it if you are selected as a presenter throughout the program.

Your capstone project is your opportunity to tell your story, know your audience and engage with them. Here are some tips to perfect your pitch.

Tip #1: The “what” matters just as much as the “why.” Providing context, setting the stage, using descriptive language, and making the “why” easy to understand the context makes the story relatable and memorable.

Tip #2: During your speech, how you speak matters just as much as what you’re saying. The tone of your voice speaks volumes. You’ll want to be strategic in how you’re emphasizing what you’re saying by the inflections in your voice. You want to record yourself to assess the type of speaker that you are and watch the recording and take notes. It’s very obvious when someone is confident and clear, and it’s also just as obvious when a person isn’t.

Tip #3: Craft a strong ending. Every story has a beginning, middle, and end. If someone only remembers how you ended, that will be the lasting impact you leave with that person.

Will your ending leave the audience wanting more?

You want us to wish we had more information and hiring partners want to reach out to you and have more questions about your project or your process and how you came up with it.

Tip #4: Your pitch should be focused. In your capstone project proposal interview, you have three minutes to pitch your project, which goes by quickly. Make sure that your main message is easy to understand, make sure that it is interesting.

Do you leave us wanting more information about your project and your processes? Were you likable during that conversation? Was it easy to understand, easy to follow? Were you confident? Those are all very important things to keep in mind.

What’s next? Learn more about the entire application process including the interview.

We’re always here to guide you through your data journey! Contact our admissions team if you have questions on how to hit your capstone project out of the park!

Related Blog Posts

Moving From Mechanical Engineering to Data Science

Moving From Mechanical Engineering to Data Science

Mechanical engineering and data science may appear vastly different on the surface. Mechanical engineers create physical machines, while data scientists deal with abstract concepts like algorithms and machine learning. Nonetheless, transitioning from mechanical engineering to data science is a feasible path, as explained in this blog.

Read More »
Data Engineering Project

What Does a Data Engineering Project Look Like?

It’s time to talk about the different data engineering projects you might work on as you enter the exciting world of data. You can add these projects to your portfolio and show the best ones to future employers. Remember, the world’s most successful engineers all started where you are now.

Read More »
open ai

AI Prompt Examples for Data Scientists to Use in 2023

Artificial intelligence (AI) isn’t going to steal your data scientist job! Instead, AI tools like ChatGPT can automate some of the more mundane tasks in your future career, saving you time and energy. To make life easier, here are some data science prompts to get you started.

Read More »