Data Browser

Code Ocean

The Product

Code Ocean empowers computational scientists to run, reproduce, and share analyses. As projects scaled, so did complexity: dozens of Datasets and were created.

The goal is to reduce time-to-find and time-to-validate datasets while increasing confidence in reuse.

The Process

The Challenge

Description-heavy rows kill scanability

High context-switching cost

No metadata hierarchy or comparison layer

Weak refinement model

“I feel I spend too much time jumping between pages to find the right dataset for my project.”

— Data Scientist, Bioinformatics Team

The Research

Competitive Benchmarking

User Interviews

Heuristic Analysis

Findings

Metadata-driven filtering is standard; free-text browsing is secondary
Faceted filters with counts reduce noise and speed narrowing
Table-based comparison enables faster triage than description-heavy lists
Saved views, cohorts, and reusable selections accelerate workflows
Provenance is part of discovery, not just audit/compliance
Context switching (open → back → open) slows validation loops
Scalable systems support column customization and bulk actions

Takeaways & Goals

Table-based > description-heavy lists

Progressive Disclossure

Provenance is part of discovery

Context switching

The Design

Principles

Architecture

Disovery/Browse Layer

Details Layer

Solution Overview

Visual Vocabulary
Calm Overview + Expandable Depth

Role -Friendly
Real-Time Synchronization

Validation & Outcomes

Results from Usability Sessions

60% ↑

Faster time-to-identify dataset

25% ↑

Increase in “In use” assets

Impact aligns user value (speed + trust), with business value (activation + reuse).

Reflection

This project shifted my thinking from “cleaning up a list UI” to redesigning a discovery system.

The real problem wasn’t visual noise, it was workflow friction.
By reframing discovery as a structured loop (refine → compare → commit), the solution became architectural rather than cosmetic.

The biggest insight: in scientific platforms, trust signals are not secondary, they are part of relevance.

This project reinforced that in complex systems, clarity is not about reducing information, it’s about structuring it so users can move with speed, confidence, and intent.