Project Overview 👋
PenguBuddy is an interactive health tracking system designed to improve the well-being of aquarium penguins through AI-powered insights and real-time behavior monitoring. Built in collaboration with Georgia Aquarium and Georgia Tech’s ACI Lab, the system combines an underwater robot and a digital platform to support both animal care and visitor engagement.
My main contribution focused on the B2B web platform for animal care staff, where I designed tools to monitor individual penguin health, flag abnormal behavior, and streamline team communication. The platform empowers caretakers to make faster, data-informed decisions while improving daily workflows.
Timeline
3 months
Winter 2024
Team
My Role
Product Designers (2B+2C)
ID Designer
Project Manager
Video Editor
Product Designer
(2B Platform)
Tool
Figma
FigJam
Notion
Context
Ensuring penguin health in zoos requires ongoing monitoring, yet traditional methods are fragmented and inefficient. Data collection is often limited, ad-hoc, and biased by time constraints. Captive penguins display just 2–23% of the activity levels observed in the wild (over 80%), highlighting a clear welfare concern.
To address this, we partnered with Georgia Aquarium and Georgia Tech's ACI Lab to develop PenguBuddy, a smart platform streamlining health tracking for data-driven animal care.
Solution
Introducing PenguBuddy
Connects visitors, penguins, and caretakers through an interactive system.
Platform Features
👩💻
Centralized Dashboard & Team Collaboration
Monitor real-time penguin health and behavior, stream live footage, and collaborate with team seamlessly. All in one smart, connected platform.
🎥
AI-Powered Behavior Alerts
📊
Data-driven Insights
Track individual penguin health in detail and manage the monitor robot.
📆
Smarter Care Planning
Manage penguin health routines with AI-suggested treatments, visual scheduling, and real-time task tracking.
AI flags unusual penguin behaviors and delivers instant video clips to support early action and informed care.
Project Timeline
Problem Space
Understand the Use Cases
Understand the Problems - What’s animal’s Welfare?
Understand the Problems - Monitoring Welfare
Through Biometrics Indicators
Through Behavioral Indicators
On-field Research
Understand the Problems - Monitoring Problems
Market Research
According to keepers…
Understand the Problems - How to improve Welfare?
2. Concept Development
PenguBuddy acts as both a welfare monitor and an enrichment tool for penguins in captivity.
By combining AI-powered real-time monitoring with interactive engagement, it not only detects early signs of health or behavioral issues but also addresses them at the root.
System Diagram
To address welfare from both enriching and monitoring system, PenguBuddy is formed by three products: an underwater robot that serves both as a toy for the penguins and a monitor, a mobile app that allows visitors to interact with the robot, and a SaaS platform that serves keepers by tracking data and automatic analysis.
3. In-depth Research
My main responsibility was designing the B2B platform. I conducted the research to define the system’s structure and data needs.
To guide this, I referenced the Five Domains Model: nutrition, environment, health, behavior, and mental state, which offers a science-based framework for assessing penguin welfare and helps keepers monitor well-being more accurately.
Current User Flow
The team currently uses a platform named ZIMS by Species 360 to track and document health data of African penguins. It is a widely used software platform designed to support the comprehensive management of animal records, medical histories, husbandry data and studbooks.
Painpoints of Current System
Stakeholder Insights
To uncover specific user needs, we interviewed various stakeholders at the aquarium to understand their experiences with the current management system, identify pain points, and gather expectations for future improvements.
Problem Statement
4. Designing MVP
Scoping the MVP
After comprehensive evaluation with cross-functional teams based on data importance, feasibility, budget, product vision, etc., I defined key data types and collection methods for our MVP.
User Flow & Wireframes
I mapped user flows based on the real-world care tasks and translated them into wireframes, aligning each screen with relevant data inputs.
Design System
I created a scalable design system that promotes visual consistency and accessibility by integrating WCAG guidelines, inclusive design principles, and clear documentation to support cross-team implementation.
Design Iteration
After an internal design audit and user testing, we identified several issues in the individual page of the initial version:
The alert section had redundant components, both a badge and an alert message.
Alert messages and follow-up recommendations, part of a continuous action flow, were siloed, disrupting the user workflow.
The design lacked flexibility, with no space for keepers to add notes that couldn’t be captured by the robot.
Using a slide-in modal limited deeper interactions like tagging or editing, leading to overlapping modals and poor usability.
Insights from a nutrition specialist highlighted feeding times as a critical data point that needed to be included.
In the updated version:
I replaced the modal with a dedicated subpage and added a breadcrumb for easier navigation.
Alert and follow-up recommendations were combined into a streamlined insight component aligned with user workflows.
I added a notes section for keepers to manually input observations, offering better flexibility.
Based on feedback from the nutrition team, I integrated feeding data into the profile and reorganized the layout accordingly.
5. Hi-fi Prototype
I. A Centralized & Semi-automated System
Centralized Dashboard & Team Collaboration
Track real-time penguin health and behavior, stream live footage, and coordinate seamlessly, all within one smart, connected platform.
Data-Driven Insights
Monitor detailed health metrics for each penguin to support informed, proactive care.
AI-Summarized Labels for Individuals
AI extracts key points from notes and generates color-coded labels by severity, making critical information easier to scan at a glance.
II. Real-time & Multimodal Data Collection
Multimodal Data Powered by Smart Robot
The robot captures video and audio of penguins, using AI to generate highlight clips that flag potential health issues. This enables more comprehensive, context-rich reviews. Staff can also manage the robot directly through the platform.
III. Streamline the Daily Care Routine
Activity Scheduling & AI-Suggested Care
Supports both calendar and Kanban views to help staff organize daily care routines. Under the scheduling section, AI suggests relevant activities, such as health checks, which can be scheduled with one click.
2 Views: Calendar and Kanban views
Schedule from AI suggestions with one click
6. Conclusion
Next Step
1. User Testing
Validate the system with real stakeholders and refine based on feedback.
2. Incorporate Environmental Data
Add inputs for factors like temperature, light, and noise.
3. Collaborate with Dev Team
Work closely with engineers to align on feasibility and implementation.
4. Scale to More Taxa
Adapt the system to support other species beyond penguins.
Potential Impact
1. Less Reliance on Manual Monitoring
Validate the system with real stakeholders and refine based on feedback.
2. Proactive Health Care
AI flags risks early, supporting timely interventions.
3. Efficient Teamwork
Centralized tools streamline daily care and communication.
4. Enrichment Through Play
Robots keep penguins active and engaged.
5. Scalable System
Designed to expand across species and facilities.
6. Long-Term Welfare Insights
Trends reveal chronic issues and guide better care.
What I Learned…
Building for Multi-Stakeholder Needs
From vets to trainers, I balanced overlapping roles and data needs to create a platform that works across disciplines.
Scalability Through Focus
Designing deeply for one species (penguins) helped me build a flexible, scalable framework for broader B2B applications.
Driving Clarity Through Systems Thinking
Simplifying health metrics, task flows, and AI outputs taught me how to turn fragmented data into actionable, user-centered systems.
Thank you for your attention 🩵
