MusiSkate improves the experience of skating through rich audio feedback.

  • Context: Class Project
  • Role: Research, Design & Management
  • Members:
    • Xiaowei Chen (Industrial Design)
    • Sarthak Ghosh (Development)
    • Pratik Shah (Development) 

Our goal is to augment the sensations of skating but not get in the way of doing tricks. That means no distracting screens or fiddling around with buttons.

We also found out that many skaters listen to music to keep them going. With MusiSkate, we combine these values of simplicity and self-motivation to design a product that enhances the learning process.

The project was presented at the Mobile HCI Conference in 2016.


Two modes

MusiSkate currently has two iterations - one using an Arduino board, and the other an Android app.


  • Unlocks a new musical track (drums, guitars, vocals) for each successful trick
  • Android app that reads accelerometer and gyroscope data


  • Provides sound effects to skating motions (jumping, 180-degree turning, swerving)
  • Uses Arduino & IMU (Inertial Measurement Unit)

Usability test session that demonstrates MusiSkate in action


Design Process

Problem discovery

In the beginning, my team was interested in skating as a vibrant subculture, but we didn't have a clear idea of what kinds of problems skaters face. And so we investigated: we went around skate parks in Atlanta, watched videos online and went through internet forums to discover common issues.

We ended up talking to a total of 11 skaters -- from 10-year-olds to 50-year-olds, students to punks -- about how they got into skating, what they enjoy about it and the issues they experience around their gear, environments and social circles.

As the most experienced researcher in the team, I led the process of creating discussion guides, conducting interviews, guiding the analysis of data, while ensuring people get experience in the various roles.


From this exploration, we discovered that skating is:

  • Simple: Less about gadgets and trends but more about mastery of skills and robust equipment.
  • Hard: Trial-and-error, repetition and research are all part of learning.
  • Social: Involves friends, families and even strangers at parks.
  • Personal: Customizing skateboards as a way to assert individuality and preferences.
  • Musical: Listening to personal music as a norm.

Informed ideation session

Based on our findings, our team got together to brainstorm on solutions we can introduce to this space. I led the structure for an organized ideation session to combine and prioritize ideas.

Direction 1: Augmented Reality

Direction 2: Musical Feedback

Direction 3: Arcade Simulation

In the end, we selected the second direction based on our common interest in music, feasibility of our concept for the duration of the course, and our professors' and classmates' feedback at a poster session.

Our model on how musical feedback (in red) can encourage the learning process.

Our model on how musical feedback (in red) can encourage the learning process.


Low to high-fidelity prototypes

Wizard of Oz prototype

We decided to split our concept into two devices so we can save time. As a low-fidelity proof of concept within our team, I designed a Wizard of Oz prototype to propose the possible user interaction and audio tracks.


Sarthak and Pratik did the heavy lifting and leveraged their own electrical engineering and computer science backgrounds to implement the two modes on Android and Arduino, while Ivy designed the task flows.


Early problem-solving through heuristic evaluation

I took charge of planning the heuristic evaluation and designing the protocols, templates and checklists, while others handled recruitment, prototyping and documentation.

Before we tested with users, another group in our HCI class conducted an evaluation of our system based on Nielsen's 10 Usability Heuristics.  The most critical issue they identified was with our music selection: Tricks for Tracks sometimes didn't give audio feedback when a trick was performed. Because of this we were able to change our song selection and prevented a disastrous outcome.


User Testing


Finally, at a parking lot on campus, we tested our system with users to see if it increases the satisfaction of skating and to compare the two modes of MusiSkate. 

Before and after testing, we had three Georgia Tech skaters rate the prototype based on: satisfaction, usefulness, appropriateness and future use.  We also conducted semi-structured interviews with the participants so they can explain their answers.

IMGP3994 (1).JPG

Our key takeaways were:

  • Audio feedback is perceived a positive source of encouragement (rewards system)
  • Music should be customizable to suit personal preferences or playlists, e.g., Spotify
  • The system must be able to produce different sounds for different tricks.
  • Freestyle is good for repetition and improvisation. Tricks for Tracks may be useful once skaters have gained mastery of tricks.

As for future points of improvement we identified:

  • Consolidation: Can we combine both prototypes into one device?
  • Modularity: How can we easily attach our prototype to any type of skateboard?
  • Sociality: How do we leverage the competitive and cooperative aspects of skating?
  • Higher rewards: How can the system detect and reward more complex tricks through audio effects or gamification?
  • Robustness: How can we improve the aesthetics of our system and prevent it from getting damaged?


Overall, this project took a lot of effort on our part, but it led to something we were collectively proud of. It was great to have a mix of team members with diverse, complementing areas of expertise, but most importantly, we had synergy because of similar work ethic and passion for our topic.

With any project, it's good to evaluate not just the design, but internal processes, communication channels and conflict resolution. Always clarify and always try to involve people even when it's not their role.

Finally, it helps to be very structured and to think far ahead and have contingency plans (e.g., hardware failing during testing). I think this is where I contributed a lot because of my UX and product development experiences. We had minimal problems and maximum preparedness, and we didn't flip-flop on a lot of our decisions.

In the end, we not only made a prototype that we all liked, but also one with tremendous potentials in the future.