Projects
Here is a selection of projects that showcase my approach to design, data, and AI. They highlight how I combine research, creativity, and technical skills to create solutions that are both functional and meaningful.
UX & Design
Credit Buddy Management App
Research-driven UX design project focused on improving financial literacy and credit management for young adults.
Figma · UX Research · Prototyping · Usability Testing
Context
Credit Buddy was developed as part of CSC396, a studio-based course centered on iterative, user-centered design. The project was completed in a team of five and progressed from early research to a fully tested high-fidelity prototype.
Research & Problem Definition
The project began with individual literature reviews examining financial literacy, credit behavior, and systemic barriers to understanding credit. These findings were synthesized to identify a core problem: young adults lack accessible and trustworthy tools for managing credit.
Primary research was conducted with a defined and reachable target audience through interviews and surveys. Results were analyzed collaboratively to extract key themes, pain points, and behavioral insights, which informed the design guidelines.
Design & Iteration
Each team member independently created low-fidelity prototypes, which were informally usability tested. Insights from testing were combined to develop a unified high-fidelity prototype in Figma.
Formal usability testing was conducted on the high-fidelity prototype, and the design was iterated based on findings related to clarity, terminology, and user trust.
Deliverables
- Research summaries and synthesized insights
- Low- and high-fidelity prototypes
- Usability testing reports and iteration notes
- Final design documentation and presentation
Adora Clothing Website Prototype
A brand and website prototype inspired by my Depop shop, focused on translating personal aesthetic and product curation into a cohesive digital retail experience.
Figma · Visual Design · Branding · UX Design
Context & Goal
Adora is a conceptual clothing brand developed to explore how a personal resale business can be translated into a cohesive online retail experience. The project focused on defining a clear brand identity, visual language, and user flow for an independent fashion website.
Brand Exploration
The goal was to articulate a consistent aesthetic that reflected the curated, vintage-inspired nature of the clothing. I explored typography, color palettes, layout hierarchy, and editorial-style imagery to create a brand that felt intentional and cohesive rather than purely transactional.
Design Process
- Defined brand values and visual direction based on existing product curation
- Designed homepage, product listings, and editorial-style sections
- Used real product images from my Depop shop to ground the prototype
- Incorporated placeholder lifestyle imagery to communicate mood and atmosphere
Outcome
The final deliverable is a high-fidelity website prototype that demonstrates branding, visual design, and layout decisions for a small-scale fashion brand, with an emphasis on storytelling and aesthetic cohesion.
Note: Product images are from my Depop shop. Lifestyle and banner imagery was used as visual reference material to communicate design intent.
Data & Analytics
TTC Subway Delays Analysis
Data analysis project examining patterns, causes, and impacts of subway delays across Toronto’s TTC network using multi-year incident data.
Python · Pandas · NumPy · Matplotlib · Scikit-learn · Data Analysis
This project analyzed TTC subway delay incidents from 2020–2024 to identify systemic patterns in delay frequency, duration, and causes. The analysis focused on temporal trends, station-level disruption, and the role of passenger-related incidents in overall service reliability.
Data Preparation & Analysis
The dataset was cleaned and standardized to resolve inconsistent station names, timestamps, and delay codes. Delay causes were grouped into passenger-related, system-related, and fire/weather-related categories to support comparative analysis across time and locations.
Key Findings
Results showed a steady increase in total delay minutes over time, with passenger-related incidents accounting for the majority of disruptions. Major transfer stations experienced disproportionately high delay durations, and late-day hours contributed most to cumulative service impact.
Modeling & Deliverables
A Random Forest classification model was developed to predict delay cause categories using temporal features and delay duration. Deliverables included cleaned datasets, exploratory visualizations, a predictive model, and a written analytical report summarizing findings and recommendations.
UFO Sightings Interactive Visualization
Developed an interactive web-based visualization of UFO sightings across Canada, focusing on data storytelling and user interactivity using D3.js.
HTML · CSS · D3.js · Data Visualization · Interactive Design
Part of CSC316, a course centered on web-based data visualization and interactive storytelling. The project allows users to explore temporal, geographic, and shape-based trends in UFO sightings, with multiple interactive features.
Research & Problem Definition
Sourced and cleaned UFO sighting reports from NUFORC. Created supplemental datasets for airports, military bases, and sci-fi films. Peer interviews and testing informed usability and interaction design.
Interactive Features
- Temporal Trends Visualization: Animated line chart showing sightings over time with year-by-year dots and pop culture markers.
- Clock Visualization: Explore sightings by time-of-day with synchronized line charts and seasonal filtering.
- Interactive Map: Key sightings vs all sightings with proximity to airports/military bases; clickable icons show details.
- Witness & Ship Count Visualization: Interactive landing zone with visual representation of witness and UFO ship counts; dynamic filtering options.
- Shape Visualization: Starry night layout showing reported UFO shapes; click to reveal descriptions and counts; toggle resizing by frequency.
Design & Iteration
Layouts and interactions were iteratively improved based on feedback to optimize readability, clarity of insights, and user engagement.
Deliverables
Complete HTML, CSS, and JS code; cleaned datasets; documentation of design decisions and iteration notes; live website and screencast.
AI / ML & Research
Employee Absenteeism Prediction Models
Built and evaluated classification, regression, and clustering models to predict employee absenteeism hours. Focused on model evaluation, feature selection, and interpretability.
Python · TensorFlow · Scikit-learn · Machine Learning
Chair Drawings: AI vs Artist
Comparative study analyzing AI-generated chair drawings versus human artists. Explored cognition, creativity, and style differences using computational methods.
Python · Machine Learning · Digital Humanities · Analysis