Spatial Intelligence

RoleLead Product Designer

CompanyEstater

ScopeUX · Interaction Design · Data Visualization

FocusGIS UX · Spatial Exploration · Decision-Making

Transformed a data-heavy reporting tool into a map-driven intelligence platform — introducing interactive layers, geospatial exploration, and flexible map-panel workflows that changed how enterprise users see and make decisions with real estate data.

01 — Context

Real estate is inherently spatial. The product wasn't.

When I inherited the Markets UX, the platform had no map experience at all. Users navigated entirely through tables, charts, and filtered reports — analytical tools that answered what, but never where.

For real estate professionals, this is a fundamental gap. A transaction price means almost nothing without geographic context. A price-per-sqft figure is abstract until you can see which neighborhoods are high, which are flat, and where the edges are. The product was communicating data without communicating reality.

02 — The Problem

Without location context, insights stayed abstract.

The absence of spatial interaction wasn't just a feature gap — it was a thinking gap. Users couldn't pattern-match across geography. They couldn't compare regions intuitively. They had to build mental models that the product should have been building for them.

No geographic pattern recognitionUsers could see that transaction volumes were high — but not where. Spotting a cluster, a boundary, or a corridor meant manually cross-referencing area codes across multiple filtered tables.

No way to compare regions at a glanceJudging one neighbourhood against another required pulling several reports and holding the numbers side by side in your head. The product couldn't show relative intensity across space.

Users had to build the map in their headsThe mental model the product should have been drawing — which areas are hot, which are flat, where the edges are — was left entirely to the user to assemble from tables.

Data without realityA price-per-sqft figure is abstract until you can see it on the ground. The product communicated numbers without the geography that gives them meaning.

03 — Design Challenge

"Add a map" was the brief. The real problem was much harder.

Adding a map is straightforward. Adding a map that makes complex real estate datasets explorable, without overwhelming users, across five distinct enterprise user types, while maintaining performance under dense data loads — that's a different problem entirely.

04 — Design Strategy

Three principles that kept the map from becoming noise

Spatial data tools fail in one of two ways: they're too simple to be useful, or too complex to be approachable. The design strategy for this project was built around avoiding both failure modes simultaneously.

01 · Start clear, add context on demandThe map opens on a clean base view; users layer on complexity one dimension at a time. Nothing is shown until it's asked for, so the canvas never becomes noise.

02 · Design backward from the decisionEvery choice started from the professional decision the user was making — a valuation, a site call — not from the dataset. The map exists to accelerate that decision, not to display data.

03 · One product, every skill levelFrom a GIS analyst who lives in maps to a developer who thinks in square footage, the interface had to feel native to all of them — so progressive disclosure teaches as you go, with no tutorial required.

The Decision I Owned

The brief was "add a map." I pushed back: a map bolted onto a reporting tool would just add clutter. I made the case to treat it as a spatial-thinking problem — an integrated map, filter, and panel loop with progressive disclosure that served five very different user types without a single onboarding screen. That reframing is what made it the product's #1 differentiator instead of just another feature.

05 — Key Design Solutions

Five solutions that turned a reporting tool into a spatial intelligence platform

SOLUTION A

Map + Panel Dual Interface

Introduced a three-mode flexible layout that respected how different users actually think. Rather than forcing a single workflow, the interface adapted to the task.

↑ Both geographic and analytical users served natively

SOLUTION B

Interactive Map Layers

Introduced a configurable layer system that let users build up geographic context incrementally — starting with a clear base view and adding analytical layers one at a time. Each layer reveals a different dimension of the market.

Avg. 2.8 layers active per enterprise session

SOLUTION C

Visual Data Interpretation

Moved data from table rows into the map canvas itself — visualizing price intensity, transaction density, and zoning status as spatial phenomena rather than numerical records. The insight doesn't require interpretation; it's immediately visible.

Insight time: 30–45 min → under 5 min

SOLUTION D

Map + Filter Integration

Connected the map and the filter/data panel into a single exploration loop — so neither component felt like a separate tool. Filters didn't just narrow a report; they updated the map in real time. And map interactions didn't just zoom in; they surfaced relevant data in the panel automatically.

↑ Exploration flow — no more context-switching between map and reports

SOLUTION E

Progressive Complexity — Designed for Multiple Skill Levels

Five enterprise user types span a wide range of spatial literacy — from government analysts who use GIS tools daily to developers who primarily think in square footage and ROI. The design was built to be immediately useful to all of them, without requiring a tutorial or onboarding sequence.

Zero onboarding required across all three user levels — progressive disclosure did the teaching

06 — Business Impact

The most transformative UX improvement in Markets' product history

The spatial intelligence layer didn't just improve an existing feature — it added an entirely new mode of product interaction that users hadn't been able to imagine before they saw it.

65%Of active sessions use map view (from 0%)

9×Faster time to geographic insight

2.8Avg. layers active per enterprise session

#1Cited differentiator in enterprise sales calls

Passive reporting became active exploration

Users stopped consuming pre-built reports and started investigating. The map interface invited a different mode of engagement — one where users drove the analysis rather than reading the output of someone else's decisions about what to show them.

Decision-making speed improved substantially

For valuation and lending decisions that previously required extended table analysis, spatial visualization cut the cognitive work of geographic understanding from 30–45 minutes to under 5. For high-volume enterprise users, that multiplies across dozens of decisions per week.

Became the product's primary market differentiator

In enterprise sales conversations, the map intelligence layer became the most frequently cited reason for selecting Markets over alternatives. Competitors offered tables and charts — this was spatial context. That distinction mattered at the procurement level.

Foundation for future monetized data layers

The layer architecture was designed to be extensible. New data layers — planning authority data, infrastructure proximity, ownership density — could be added and monetized independently, creating a path to new revenue streams that didn't exist in the original product structure.

07 — Why This Matters

GIS UX at enterprise scale is a rare skill. Most portfolios don't show it.

Spatial interface design is a distinct discipline — it combines information architecture, data visualization, interaction design, and performance UX in a domain where bad decisions have real cognitive consequences for high-stakes professional workflows.

GIS & Spatial UX🌍 Designing for map-based interaction requires understanding how people read spatial data — which is fundamentally different from how they read dashboards or tables.

Multi-layer Information Architecture📐 Managing visual complexity across simultaneous data layers without overwhelming users is an IA challenge as much as it's a design one.

Progressive Disclosure at Scale🎛 Making the same product feel right for a GIS novice and a professional spatial analyst requires thinking about disclosure as an ongoing interaction, not just an onboarding decision.

Interaction Design at the System Level🔗 The filter-map-panel loop required designing how three distinct components communicated — not just how any one of them looked or behaved in isolation.

Performance-Conscious Design⚡ Map rendering under dense data loads means design choices have engineering consequences. Layer prioritization, progressive loading, and simplified default states were all performance decisions as much as UX ones.

Decision Workflow Design🎯 The goal was never to display data. It was to accelerate professional decisions. Designing backward from the decision — not forward from the dataset — is what made this work.

08 — Challenges

What made this technically and conceptually hard

C1 · Avoiding visual overload with dense datasetsReal estate transaction data is dense and irregular — clustered in cities, sparse in outskirts. Designing visualizations that remained legible at all zoom levels, under all data densities, required close collaboration with the engineering and data teams on rendering thresholds and aggregation strategies.

C2 · Managing multiple simultaneous layers without confusionEach additional layer added visual complexity. The risk was that two or more layers together would become unreadable — or worse, appear to show more information than they actually did. Color, opacity, and draw order had to be designed as a system, not individually.

C3 · Performance with real geographic data at scaleMap rendering under production data loads exposed performance issues that affected design decisions — layer complexity limits, progressive loading states, and zoom-level data aggregation all required design responses, not just engineering ones.

C4 · Educating users on new interaction patternsThe layer system and map-panel loop were genuinely new interaction models for most users. Getting users to discover and trust non-obvious behaviors — without an onboarding wizard — required careful progressive disclosure design and thoughtful default states.

09 — What I Learned

Visual context doesn't simplify data. It changes how you think about it.

When dealing with complex data, visual context reduces cognitive load more than simplification alone. Maps don't just display data — they change the kind of thinking that data enables.

The most unexpected outcome of this project was watching users ask completely different questions once the map was in their hands. Before, they asked "what are the prices in Zone 4A?" After, they asked "why does that corridor have higher intensity than the areas around it?" The visualization didn't just speed up existing analysis — it prompted analysis that wasn't happening before.

That's the deepest version of what good data UX can do: not just surface what's already known, but create the conditions for users to discover something they didn't know to look for.

More Projects

Estater × Power BI

BI Integration & Systems Design · Estater

Design Leadership

From Solo Designer to Design Function · Estater