County view
A dense overview keeps the full record set visible while the interface explains what each mode changes.
The review path
What to notice before you click.
The live app is intentionally dense. This is the shortest path through the experience: county context, search, focus, semantic neighborhood, and map handoff.
1. Start with county context
All 8,406 records stay visible so the search result has a place in the larger network.
2. Search for a need
Typing “coffee” brings the most relevant businesses forward without reducing the demo to a flat table.
3. Enter a neighborhood
The selected business becomes an anchor, nearby businesses glow, and the lines show semantic connections.
4. Open the map
The focused record carries into terrain mode, then returns to the network without losing context.
Working surfaces
Built as a real interactive system.
The point is not a static chart. The explorer has separate states for full-network browsing, search focus, neighborhood walking, and map return.
Similarity walk
Focused neighborhoods make semantic relationships inspectable instead of leaving the graph as a decorative cluster.
Map handoff
The selected business can land on terrain and return to the network, preserving the visitor’s place.
Technical scope
What this exercised.
This is a portfolio artifact for dense data UX: semantic search, browser rendering, state choreography, mobile hardening, and live deployment discipline.
- Rendered a high-density local business graph with responsive overlays and touch-safe controls.
- Separated exploration modes so website/contact signals, connector records, and semantic neighborhoods are understandable to a first-time visitor.
- Hardened the live interface with visual, control, runtime, and small-screen layout checks before presenting it publicly.
- Kept the public story honest: this is a working technical demo over real local records, not a sales page.
The core achievement is simple: public business records become a navigable network, and the interface keeps the visitor oriented while moving between search, graph, neighborhood, and map.
Stack: HTML, CSS, JavaScript, Three.js, Leaflet, public record data, and generated semantic thread artifacts.
Under the hood
The visual layer sits on a data pipeline.
The explorer is the public surface of a LeadOps workflow: local records are cleaned, public-safe fields are exported, semantic artifacts are generated, and the browser renders those artifacts as searchable neighborhoods.
- LeadOps keeps the raw business corpus, research notes, and audit signals separate from the public demo export.
- The public artifact exposes safe names, locations, categories, summaries, and semantic neighbors for 8,406 records.
- Generated thread files give each record nearby candidates so the app can show relationships instead of isolated search results.
- Three.js handles the dense graph, Leaflet handles terrain handoff, and the UI keeps state when visitors move between modes.
The technical takeaway: this is not just a polished front end. It is a shipped interface over a structured data workflow with public/private boundaries, generated semantic context, and browser-level QA.
Proof path: LeadOps corpus → public-safe export → semantic search and thread artifacts → live WebGL explorer.
Try the live explorer.
Start in County View, search for a familiar need like “coffee” or “roof repair,” then enter the neighborhood around any result.