Professor Duri Long Grace Wang Hasti DarabiPourshiraz Nyssa Shahdadpuri
What is DataBites?
DataBites in Griffin Museum of Science and Industry 2025
DataBites is an interactive exhibit that introduces children to the basics of machine learning in a playful way. Kids build their own dataset, learning that data quality and variety matter. Then, they train an AI model with this dataset, then test its performance. And as the AI classifies new data, children see firsthand how results change depending on whether their dataset was complete, biased, or representative. This hands-on process gives children a tangible understanding of how machines “learn” and why data matters.
My Role / Project Goal
NewDataBites Design I Produced & Implemented
Based on insights from UX and learning science research, I cooperated with the researchers to produced comprehensive UX and visual design update and visual
The goal was to strengthen the learning outcomes through clear delivery of AI education, emphasizing
Previous DataBites? - UX Research Recap
PreviousDataBites
Design Jam in Collaboration with Georgia Tech's TILES Lab
(I joined the lab in Summer 2024 after the study was conducted but contributed by segmenting and analyzing the video data to uncover insights for design improvements. I then participated in discussions on how to apply these insights to enhance DataBites’s learning outcomes and user experience.)
Setting?
2-day study at the Griffin Museum of Science and Industry (Chicago), Summer 2024
Participants?
Children aged 10–14, recruited on-site
Data Collected?
- 95 segmented video recordings - Live observations - Structured interviews
Focus of Analysis?
- Active, prolonged engagement - Understanding of AI concepts - Situational interest - Creativity and embodiment
1. Learning Experience in the Physical-Digital UX
👍 Using tangible materials (e.g., building pizza and sandwich datasets) encouraged prolonged, playful engagement.
👎 The UI/UX created bottlenecks, as visitors often didn’t know where to focus next. The scattered setup of buttons, projection, and dataset creation table made the process confusing.
➡ Rearranging the Exhibit Setup
2. Machine Learning Concept Delivery
👍 Visitors enjoyed creatively building their own datasets and observing how this influenced results.
👎 Explanations of the process and outcomes were lacking, and the connection to machine learning concepts was weak.
➡ Emphasizing AI in a Playful Way
➡ Visualizing the Machine Learning Process More Clearly
➡ Guiding Visitors through Step-Focused Buttons
➡ Integrating More Interactive Engagement in the Machine Learning Process
New DataBites? - Key Design Updates
1. Strengthening the Learning Experience in the Physical-Digital UX
1.1 Re-arranging the Exhibit Setup
Previous Setup
New Setup with Double Screens and a Circular Table
➡ The revised layout fosters more active engagement across both physical and digital interactions by allowing visitors to take in the entire exhibit from a single vantage point.
2. Clearer Delivery of Machine Learning Concepts
2.1 Emphasizing AI in a Playful Way
Introduced a food-themed AI character to guide visitors’ attention toward machine learning
Added a voice guide to reinforce focus and explain actions
➡ Highlights AI throughout the experience and connects visitors’ actions to learning outcomes
2.2 Visualizing the Machine Learning Process More Clearly
Previous Flow: Create Dataset -> See Result
Brainstorming: Identifying missing steps in the process, such as the training step
Mock-Ups: Defining and visually presenting each step of the machine learning workflow
Updated Flow: Create Dataset → Train → Create Testing Dataset → See Result
➡ Makes the machine learning development process more explicit and easier to understand
2.3 Guiding Visitors through Step-Focused Buttons
Previous buttons (Start, Train, Restart) were always visible
New buttons (Begin, Create Your Dataset, Train, Test, Try Again) appear only when relevant, with labels aligned to learning goals
➡ Reduces confusion and guides visitors through each step of the process, more effectively delivering the learning goals
2.4 Integrating More Interactive Engagement in Machine Learning Development
Previously static steps (training and testing) were made interactive: Training: visitors watch the procedure update in real time. Testing: visitors select images to test their AI.
➡ Enhances engagement and reinforces learning through hands-on participation
Final Thought
At Museum of Science and Industry 2025 with New DataBites
This project was again the result of active collaboration among back-end developers and learning science researchers, which sparked discussions about what information to present and how to structure the learning experience effectively. And it gave me the opportunity to design across both physical and digital realms, considering how visitors move through the space, how their attention shifts between physical and digital interactions, and what elements to emphasize in each domain.
Moreover, wearing multiple hats as a motion designer, visual designer, UX designer, and front-end developer allowed me to approach the project from diverse perspectives, from responsive front-end implementation to creating engaging visuals.
My favorite part of this project was when we hosted our updated DataBites at the Musem of Science and Industry. Observing firsthand how visitors explored, experimented, and learned through trial and error that we designed, I found a deep fulfillment and passion in creating works that inspire people with creativity and knowledge.
