Open Data Jabar Improving Experience

Improving dataset discoverability through homepage restructuring and predictive search redesign

Overview

Open Data Jabar is a public platform providing access to datasets, maps, and visualizations from the West Java government.

The platform serves diverse users:

• Policy makers

• Researchers

• Journalists

• Students

• General public

Despite having extensive datasets available, users struggled to efficiently discover relevant data.

My goal was to transform the platform from a static information portal into a guided discovery experience.

The Problem

1. High Scroll, Low Interaction

   Users scrolled extensively — especially on mobile — but interaction rates were low. Visibility did not translate into engagement.Users scrolled extensively — especially on mobile — but interaction rates were low. Visibility did not translate into engagement.

   
   

2. Over-Concentration on Top Sections

   Most engagement occurred only in the upper section of the homepage. Deeper content sections were rarely explored.

   
   

3. Friction in Search Flow

   The previous search experience required users to:
   
   

   1. Click search

   2. Navigate to a dedicated search page

   3. Enter a keyword

   4. Manually refine results

There were no:

• Real-time suggestions

• Trending dataset recommendations

• Related dataset surfacing

Search functioned as a utility — not as a discovery tool.

Core Challenge

Designing for a public government platform means designing for everyone
The challenge was balancing:Designing for a public government platform means designing for everyone
The challenge was balancing:

1. Simplicity for first-time users

2. Depth for experienced data analysts

3. Accessibility across mobile devices

4. Clarity in a data-heavy environment

Constraints

As a government product, the redesign operated under several constraints:

Search relied on an existing backend API

• Filtering options were limited to available dataset metadata

• Government branding guidelines restricted drastic visual changes

• Stakeholder approvals were required before release

• Limited development resources affected iteration speed

  
  

Strategy Shift

Rather than treating search as a separate feature, I reframed the platform around two strategic shifts:

1. From Passive Homepage to Discovery Hub

   • Rebuilt the homepage with modular sections

   • Elevated high-value datasets to top positions

   • Improved visual hierarchy for better scanning

   • Reduced content overload

     
     

2. From Reactive Search to Predictive Assistance

   • Enabled inline search interaction

   • Introduced real-time query suggestions

   • Surfaced trending and popular datasets

   • Displayed related datasets during search

   • Reduced unnecessary page transitions

Design Process

1. Understanding the System & Behavior

Before sketching anything, I needed to understand two things:

• How users behaved

• What the system allowed

I reviewed analytics across a 3-month window and observed:

• High scroll depth, especially on mobile

• Low interaction in deeper homepage sections

• Concentrated engagement at the top of the page

• Friction in search due to multi-step navigation

I also aligned with data analysts to understand:

• Existing search API limitations

• Available metadata fields

• Technical feasibility for real-time suggestions

2. Mapping the Current Experience

To clarify friction points, I mapped the existing flow.

Pain Points:

• Extra page transition

• No predictive support

• No exploratory guidance

• High cognitive load for new users

  
  

3. Mapping the Current Experience

Before jumping into wireframes, I defined 3 principles to guide decisions:

1. Reduce Friction

   Minimize steps and unnecessary transitions.

2. Guide Discovery

   Support users who don’t know exact dataset names.

3. Balance Simplicity

   Maintain advanced filtering without overwhelming new users.

   
   

   
   

4. Exploring Move Pixels

Start explored multiple approaches

5. Concept on How We Solve the Problems

The redesign prioritized clarity over density, ensuring users could quickly identify relevant content without excessive scrolling

Validation & Impact

The redesign was deployed in staging and validated through user testing and stakeholder review.The redesign was deployed in staging and validated through user testing and stakeholder review.

• Reduced Time-to-Find

  During usability testing, users were able to locate datasets faster compared to the previous experience.

• Improved Search Confidence

  Users responded positively to:
  
  

  • Real-time suggestions

  • Trending dataset visibility

  • Simplified interaction flow

  
  The predictive elements reduced uncertainty, especially for users unfamiliar with dataset naming conventions.

• Positive Stakeholder Feedback

  Stakeholders noted that:
  
  

  • The homepage felt more structured and purposeful

  • The search experience appeared more intuitive

  • The recommendation concept aligned with long-term engagement goals

  
  Due to internal policy, detailed production metrics cannot be publicly disclosed.

Key Learnings

1. Search Is About Guidance

   Many users don’t know exactly what they’re looking for. Predictive support improves confidence and reduces friction.

   
   

2. Engagement Requires Hierarchy

   High scroll depth does not equal meaningful interaction. Content prioritization matters more than content quantity.

   
   

3. Public Products Require Balance

   Government platforms must remain accessible while supporting complex data exploration needs.
   
   Designing for a broad audience requires simplicity without oversimplifying functionality.