Scientific App panel.
A no-code panel for scientists to run and compare computational results without relying on engineering support.
Project Summary
A reproducible no-code panel that helps scientists run, inspect, and compare computational outputs without writing scripts.
Role Overview
Owned UX strategy and delivery from problem framing to final handoff, including prototyping, visual systems, and implementation documentation.
Design Process.
How we moved from a fuzzy problem to a shipped solution.
Problem framing
Scientists depended on engineering for routine analysis
The existing workflow created bottlenecks and delayed validation loops because non-coder users lacked clear controls and understandable output states.
Project Goal
Enable independent, reproducible analysis in no-code UX
Build a guided panel with clear run states, comparison workflows, and readable metrics that reduce cognitive load.
Research & Takeaways
What users were actually saying
We mapped dependencies between scientists and engineering and identified the moments where unclear language and overloaded controls blocked adoption.
Caption: Workflow dependency map from user research
Wireframes
Exploring the solution space
Wireframes focused on action clarity across run setup, result interpretation, and comparison. We iterated on hierarchy and simplified critical decisions.
Results
What shipped and what moved
There was no way for scientists to review pipeline outputs or save results directly inside Code Ocean; that workflow lived outside the platform entirely. The App Panel changed that. Time-to-create a new dataset dropped 58%, capture events per completed run increased 41%, and support tickets around file confusion fell roughly 60% in the first six weeks. Scientists stopped leaving the platform to validate their work and started trusting it to hold it.
