Research Tools
When the right tools don't exist within my student budget, I build them. Each instrument here is fully custom and purpose-built for real UX research studies. I'm working on making that process repeatable for anyone.
TreeTest AI
The custom tree testing tool I built for ShopGoodwill evolved into TreeTest AI — a free, AI-powered IA research pipeline co-built with Drake Durflinger. It turns the methodology behind the studies on this page (card sort → site map → tree test → analysis) into a repeatable pipeline any researcher can use, regardless of budget.
Visit TreeTest AI →How I Build
Every tool starts with a research gap — a study I needed to run that no free platform could support. From there, I work iteratively with Claude and Claude Code to move from architecture decisions to working implementation, using AI as a development collaborator rather than a shortcut. Claude handles the boilerplate; I stay focused on the research logic, the participant experience, and the data structure that needs to come out the other end.
Under the Hood
The tree testing platform runs on a Firebase + Firestore backend, with a participant-facing session flow (screener, consent, tasks, SEQ, closing questionnaire) and a researcher dashboard that renders six Canvas-based visualizations in real time: scatter matrices, first-click heatmaps, box plots, SEQ line graphs, stacked bar charts, and a navigation path dendrogram.
The Research Logic
These tools are built using the research plan I created. Each data point is captured to measure, test, or uncover data points relevant to the study objective. Task structures, success scoring, directness rates, and SEQ collection are all deliberate decisions informed by IA evaluation methodology. The v1 → v2 iteration on the ShopGoodwill tree test was driven entirely by data from the first version, which is exactly what the dashboard was built to enable.
ShopGoodwill Card Sort
A hybrid card sort built to evaluate ShopGoodwill's category structure. Participants drag items into predefined sections and either use existing sub-sections or create their own, giving me raw data on how people naturally organize the site's inventory. The researcher dashboard visualizes sorting patterns with similarity matrices, dendrograms, and category agreement metrics.
ShopGoodwill Tree Test V2
The second iteration of my tree test, redesigned based on findings from V1. Participants complete shopping tasks against a revised category structure, rate difficulty after each task, and provide open-ended feedback. The researcher dashboard includes real-time analytics and CSV export.
ShopGoodwill Tree Test
The first version of my tree test, now closed with 12 completed sessions. This study established baseline findability metrics for ShopGoodwill's original navigation structure. The results dashboard surfaces success rates, navigation path analysis, and task difficulty scores that informed the V2 redesign.