TELUS Space Booking Platform

Overview

TELUS, a major Canadian telecommunications provider, faced a significant challenge managing their vast commercial real estate portfolio. With over 1,000 offices nationwide, their wealth of building data was locked away in REVIT models and a legacy Computer Aided Facility Management (CAFM) system. This created a bottleneck in providing employees with modern workplace amenities like desk booking and navigation.

I was tasked with unlocking this data to create an intuitive desk booking solution that would leverage their existing building information models while delivering a seamless employee experience.

Project Goals

As Technical Lead for this project, I managed a team of 4 developers focused on transforming complex building data into an accessible desk booking platform. Here were the requirements:

• Every real estate property should have a dynamically generated floor plan that is interactive and allows users to book desks.
• These interactive maps must be able to performantly render on web and mobile devices.
• The maps must be editable by administrators to allow for specific furniture and equipment to be added.
• The maps must be rendered in 2D.
• The maps must be installable as an NPM package into any TELUS React project.

My Role

I was the Technical Lead for this product and I worked closely with the Product Owner to define the requirements and the technical approach. As the Technical Lead, I was responsible for designing solutions to each problem we faced and implementing them.

Technologies Used

• React
• Next.js
• PixiJS
• TypeScript
• Neo4J

Challenges and Solutions

Converting REVIT Models to Interactive Maps

"How do we convert complex REVIT models into lightweight, interactive floor plans that can be rendered on any device?"

Our first challenge was extracting just the floor plan data we needed from the complex REVIT models, while keeping everything spatially accurate.

Our solution was to develop a custom pipeline, which relied on our partners at TELUS to manually label the data.

1. Import REVIT models into the CAFM system
2. Manually label data by:
   • Drawing regions over bookable spaces
   • Assigning relevant metadata
   • Maintaining spatial integrity of the floor plan
3. Export labeled data into Neo4J as a graph of points
4. Query the graph database to retrieve all spaces and points for each floor
   • Even largest floors only required ~20kb of data
   • Data was cached since maps were entry point to app

This resulted in spacially accurate data with appropriate metadata for each space.

In this image you can see the data we extracted from the REVIT model, represented a set of bookable desks.

How do we render performantly?

"How do we achieve smooth performance when rendering large floor plans with thousands of interactive elements?"

We approached this challenge by treating our floor plans like an isometric video game, utilizing several game engine optimization techniques:

1. Custom Tile System

   • Commissioned custom pre-generated isometric tiles
   • Created a compressed tile sprite library
   • Tiles represented different floor plan elements (desks, rooms, etc.)

2. Custom PixiJS Rendering Engine

   • Built optimized WebGL renderer using PixiJS
   • Implemented key optimizations:
     • Viewport culling (only render visible tiles)
     • Mipmap caching for different zoom levels
     • Client-side tile caching
     • Texture atlasing to minimize draw calls

3. Performance Results

   • Achieved 60fps even on mobile devices
   • Maps with 1000+ interactive elements rendered smoothly
   • Initial load time under 2 seconds
   • Cached subsequent loads were near instant

This game engine approach allowed us to create highly performant interactive maps that worked well across all devices while maintaining visual fidelity.

How do we keep this up to date?

Map Editing System

To maintain accuracy and adapt to changing office layouts, we developed a parallel map editing system for administrators. This system allowed for real-time updates to the space configuration without requiring new REVIT exports.

Key features included:

1. Administrative Interface

   • Drag-and-drop space relocation
   • Non-bookable furniture placement
   • Visual wayfinding elements
   • Real-time preview of changes

2. Change Management

   • Version control for floor plan changes
   • Audit trail of modifications
   • Scheduled publishing of updates
   • Rollback capability for errors

3. Wayfinding Elements

   • Placement of directional signage
   • Addition of landmarks and reference points
   • Creation of suggested walking paths
   • Accessibility route markers

This editing system empowered facility managers to keep floor plans current and improve the user experience through strategic furniture placement and wayfinding elements.

Outcomes

The TELUS Space Booking Platform was successfully launched internally with significant coverage and adoption:

• Scale of Implementation
  • Mapped 300 floors across 59 buildings
  • Created 26,500 bookable spaces
  • Deployed to 10,000 active employees

This large-scale deployment demonstrated the platform's ability to handle enterprise-level requirements while maintaining performance and usability across a diverse portfolio of commercial real estate assets.