User Research Portfolio · SYDE 548, University of Waterloo
Engineering course selection at UWaterloo forces third- and fourth-year students through an unfiltered catalog, disconnected tools, and confusing feedback — with real cognitive and emotional cost. This portfolio uses course selection as the vehicle to practice a full user research toolkit, centered on cognitive load and accessibility for neurodivergent and international students.
The problem
UWaterloo engineering students can't personalize their schedules until upper years, when they finally choose technical and complementary studies electives. They're met with a long, unfiltered course catalog, many of which they're not eligible for or that won't count toward their degree. Specializations add another layer of complexity, and existing tools don't surface certificates, options, or minors that could enhance a degree. Finding courses students are passionate about, can take, and that count toward graduation has proven tiresome and error-prone.
Redesign course selection so students can efficiently choose electives while tracking degree requirements and surfacing opportunities to enhance their degree.
Goals
I split the experience into three moments, each with its own goal: students feel empowered preparing for selection, stay motivated to complete it on time, and trust that their choices fulfill their degree requirements once submitted. Usability goals followed the same structure: effective to use (≥80% of selected courses meet requirements), efficient to use (≤15 minutes to finalize a term), and satisfying to use (≥8/10 on a Likert scale).
Persona
To keep the redesign honest to real constraints, I built a proto-persona with a specific accessibility and diversity lens rather than a composite "typical" student.
Jordan Baker
22, Third-Year Mechanical Engineering, International Student
Research methods
Three participants, two fourth-year Chemical Engineering students and one proxy playing the Jordan persona, each wrote a Love Letter and a Break-Up Letter to the course selection experience, then walked through a pre-defined journey map scoring each step on a Likely-to-Recommend scale. The letters and comments were transcribed and run through an inductive thematic analysis, first with ChatGPT-4.0 as a starting pass and then refined manually against what I'd observed firsthand in the sessions.
The lowest-scoring points in the journey were checking degree requirements and confirming eligibility. Participants consistently said a visit to their academic advisor was needed to feel confident, even after using every tool available to them.
Findings
| Theme | Frequency | What it captures |
|---|---|---|
| Fragmented information ecosystem | 17 | Frustration at needing Quest, UW Flow, the Schedule of Classes, and degree checklists all at once |
| System feedback & clarity gaps | 7 | No clear feedback until submission — "what does that even mean, I don't meet the prerequisites?" |
| Poor visual & UX design | 5 | Outdated, unintuitive layouts across every tool in the process |
| Search & filter functionality | 5 | A rare bright spot — students liked alphabetized course lists |
| Communication & planning support | 4 | Email reminders were the one thing keeping students accountable |
A follow-up literature review and a modified fly-on-the-wall pass through the UWaterloo subreddit confirmed the pattern: students consistently improvise spreadsheets and third-party trackers because no single tool consolidates degree progress, and academic jargon compounds the problem for international students navigating an unfamiliar system.
Heuristic evaluation
Evaluating five tools in the process against Norman's design principles surfaced two critical failures: the Academic Calendar gives no guidance preventing students from planning against an outdated year's requirements, and the course cart doesn't surface prerequisite or scheduling conflicts until the moment of submission, by which point a student may need to restart the entire selection process.
Reflection
Narrowing scope to third- and fourth-year engineering students was a deliberate trade-off. It was broad enough to matter, specific enough to design for honestly. The proto-persona and persona brief exercises pushed me to catch my own bias in real time: my first instinct was to build accessibility considerations I was already familiar with, rather than sit with edge cases I hadn’t personally experienced. GenAI was useful for a fast first pass on thematic analysis, but it flattened UWaterloo-specific context until I fed it my own material. It was a reminder that tools built for general users need a human closing the loop on specificity.