<\!-- MOBILE MENU -->
Work About Career Contact
Oracle · 2019–2021

Oracle Unity Data Platform

Transforming data integration from complex JSON workflows into an intuitive Admin Console. By redesigning the system with Redwood style guidelines and visual configuration interfaces, we enabled System Administrators to create job configurations, ingest data, and manage schemas 70% faster—eliminating the need to work directly with raw JSON files.

Metrics & Impact

70%
Speed improvement in routine admin work via visual interface
6 Phases
Comprehensive redesign from JSON Tool Manager to Admin Console
Zero JSON
Admins no longer required to work directly with raw JSON files

Research & Strategy

The Core Challenge

The legacy MCPS-Manager exposed raw JSON data structures that System Administrators had to manually parse, understand, and manipulate. Admins like Gavin needed to create Job Configurations to ingest customer data, but the process was cumbersome: working directly with complex JSON files, managing source detection, and mapping columns—all without visual assistance.

The result: slow, error-prone workflows and significant cognitive load on a small team managing critical data integration tasks.

Design Goals

  • Eliminate direct JSON interaction for routine admin tasks
  • Enable visual job configuration creation and management
  • Provide intuitive source detection and schema mapping
  • Make data ingestion workflows discoverable and searchable

The Admin Console Solution

Redesigned the entire platform using Oracle's Redwood Style Guide, creating a visual-first Admin Console where System Administrators can:

  • Create Job Configurations through an intuitive form-based interface (no JSON editing)
  • Detect Sources and automatically display available schemas from connected data sources
  • Map Columns visually by dragging source columns to Unity Data columns
  • Search & Manage all job configurations in a clean, searchable view
  • Ingest Data with clear, step-by-step workflows that guide admins through complex processes

Key Impact

  • 70% faster routine admin workflows (job creation, data ingestion, column mapping)
  • Zero JSON files required for standard operations—all interactions visual and intuitive
  • Self-service platform enabling admins to work independently without deep technical knowledge of data structures
  • Scalable operations allowing the admin team to manage significantly more job configurations

6-Phase Redesign Approach

Phase 1: Study JSON Tool Manager

Deep analysis of the legacy MCPS-Manager system. Documented how System Administrators currently worked with JSON structures, identified pain points, and mapped the data flow from source detection to column mapping to data ingestion.

Phase 2: Navigation & Information Architecture

Designed a clear, hierarchical structure for the Admin Console. Created navigation patterns that surface key actions (Create, Search, Edit, Delete) and organized screens logically: Dashboard → Job Configurations → Configuration Details → Execution Logs.

Phase 3: Data Ingestion Design

Built intuitive workflows for data source selection and connection. Designed forms and wizards that guide admins through source detection, validation, and configuration without requiring JSON knowledge.

Phase 4: View, Edit, Create Pages

Designed comprehensive CRUD interfaces following Redwood style guidelines. Created detail pages with inline editing, confirmation dialogs, and clear error messaging. Built list views with filtering, sorting, and search to help admins manage hundreds of job configurations.

Phase 5: Rebrand Query Builder

Transformed the query builder interface from a text-based JSON editor into a visual query composer. Admins can now build complex queries through an intuitive interface with drag-and-drop column selection, operator dropdowns, and real-time query preview.

Phase 6: Shell & Integration

Implemented the unified shell/navigation framework, integrated all modules into a cohesive Admin Console, and ensured consistent styling across all components using Redwood design system.

Understanding the System Administrator

Primary User: Gavin (System Administrator)

Role: Manages data integration workflows and job configurations for Oracle Unity Data Platform.

Key Responsibilities:

  • Create and maintain Job Configurations
  • Detect and map data sources
  • Define column transformations and mappings
  • Monitor data ingestion execution
  • Troubleshoot failed imports

Pain Points Addressed

  • JSON Complexity: Gavin had to manually edit JSON, making simple tasks error-prone
  • Time Investment: Routine tasks took hours due to manual JSON manipulation
  • Learning Curve: New admins required significant training on JSON structure and syntax
  • Error Rate: Manual JSON editing led to typos, syntax errors, and failed imports
  • Discovery: Finding existing job configurations required searching through hundreds of JSON files

Key Insights

Mental Model Shift: Gavin thinks in terms of "data sources," "columns," and "jobs"—not JSON structures. The redesign centered on his mental model, not the underlying data format.

Visual Preference: Admins strongly prefer visual interfaces. When shown mockups of the visual column mapper vs. JSON editing, Gavin's team unanimously chose the visual approach.

Searchability Matters: Gavin manages 500+ job configurations. The ability to quickly search and filter was critical to his workflow.

Confidence in Decisions: When Gavin can see a preview of his configuration before saving, he makes decisions faster and with fewer revisions.

Quote from Gavin

"The old system forced me to become a JSON expert just to manage configurations. With the new console, I can focus on the data itself—which sources to connect, which columns to map—instead of wrestling with JSON syntax."

Measurable Impact

Workflow Efficiency

Before: Creating a job configuration took 45 minutes (JSON file creation, manual column mapping, syntax verification, testing).

After: Same task takes 12–15 minutes using the Admin Console (guided workflow, visual mapping, instant validation).

Result: 70% time savings per configuration

Error Reduction

JSON syntax errors dropped significantly. The visual interface with real-time validation eliminated typos and structural errors before configurations were saved. Failed imports due to configuration errors reduced by 85%.

Adoption & Scalability

New System Administrators could become productive in days instead of weeks. The team confidently manages 2.5× more job configurations with the same headcount, enabling rapid expansion of data integration capabilities across Oracle's product lines.

Strategic Impact

Self-Service Platform: System Administrators work independently without requiring data architects to write JSON code.

Faster Go-to-Market: New data source integrations launch faster when configurations can be created visually.

User Satisfaction: Gavin's team reported significantly higher job satisfaction—managing data instead of fighting JSON syntax.

The Complete User Journey

The Admin Console (Job Configs) is the critical foundation of the entire Oracle Unity Data Platform. It enables System Administrators like Gavin to efficiently ingest data, which then flows through mastering and analytics, ultimately enabling downstream users to explore and report on clean, unified data.

Phase 1: Ingest Data
System Administrator
  • ✓ Create job config
  • ✓ Configure sources
  • ✓ Map columns
  • ✓ Run publish job
Phase 2: Master Data
Instance Admin & Power Users
  • ✓ Unify data
  • ✓ Augment attributes
  • ✓ Promote records
  • ✓ Run analytics
Phase 3: Use & Explore
Data Consumers & Explorers
  • ✓ Query data
  • ✓ Generate reports
  • ✓ Export for use
  • ✓ Build insights
The Admin Console is the Gateway

Without efficient job configuration and data ingestion, the entire downstream mastering and exploration pipeline cannot succeed. The Admin UI directly enables 70% faster data integration workflows.

View detailed user journey →

Full Case Study

Download the full case study PDF for detailed design process, data architecture, and research findings.

Interested in discussing this work?

I'm always happy to discuss design process, data platform design, or how I approach making complex technical workflows approachable.

📄

Download Case Study

Enter the password to download the full PDF with wireframes and research findings.