← All work
Analytics Platform

Glasswing

Hyatt Hotels

A self-service analytics app that turns four layers of hotel performance data into a story general managers can act on.

Glasswing is a progressive web app that puts Hyatt’s business performance in the hands of the people who actually move it — corporate sponsors, analysts, owner/operators, and the general managers running individual properties.

The problem

Performance reporting at the property level ran on a legacy process that leaned on manually generated reports. They were slow to produce, frequently inaccurate, and stale by the time they reached a GM’s desk. Leaders were making operational decisions on numbers they couldn’t fully trust — and franchisees in particular had little visibility into how their property stacked up against the rest of the portfolio.

What I built

Data from a range of upstream sources is aggregated into Snowflake and exposed through ThoughtSpot, a self-service analytics platform. Glasswing sits on top of that pipeline as the experience layer: a fast, filterable app where users can interrogate the numbers on their own terms — by property, time period and comparison period, region, management group, franchised vs. corporate-owned, currency, and portfolio benchmark.

The harder design problem wasn’t access to data — it was meaning. Hyatt’s metrics fall into four categories that feed one another in sequence: guest satisfaction → brand loyalty → yield management → financial performance. Each link in that chain influences the next. Most dashboards flatten that relationship into rows of disconnected KPIs. Glasswing is built to surface the chain itself — showing how a move on an upstream metric ripples downstream — so the tool reads less like a report and more like a narrative that points toward a specific action.

Why it matters

By making the cause-and-effect visible, Glasswing aligns GMs to Hyatt’s broader operational vision and gives franchisees something they never had: a clear, trustworthy benchmark against comparable Hyatt properties and groups. It replaces an error-prone manual process with current, self-service data — and turns “here are your numbers” into “here’s what to do next.”

Stack

  • React + TypeScript — application front-end (installable PWA)
  • Node.js — backend microservices
  • ThoughtSpot — self-service data analytics platform
  • Snowflake — aggregated data warehouse
  • AWS — hosting, routing, infrastructure-as-code, and observability