Travelin

Overview & Background

Travelers often struggle with fragmented information, inefficient route planning, and manual itinerary creation.

Our team engineered Travelin, an end-to-end web platform designed to streamline trip planning, enable community itinerary sharing, and deliver real-time destination recommendations for travelers across Indonesia.

What We Did

• Context-aware recommendation engine delivering destination suggestions based on spatial proximity, user activity preferences, Google Places API data (backed by an offline fallback strategy), and crowd-sourced ratings.
• Content sharing and discovery enabling users to publish, explore, like, and duplicate itineraries as baseline templates.
• Collaborative trip planning and budgeting enabling users to plan multi-day itineraries with time-validated schedules, real-time activity cost tracking against trip budgets, and role-aware sharing permissions across public, private, and link-based access states.
• Production-Grade backend architecture Built on top of UUID primary keys, relational Eloquent models, soft and cascading deletes, multi-record database transactions, strict validation, and middleware security.

Challenges & Key Accomplishments

Spatial Proximity Optimization in Recommendations

• Challenge: Recommending travel spots solely on ratings often led to geographically impractical suggestions scattered across distant regions.

• Solution: Integrated spatial coordinate mapping via the Google Places API alongside the Haversine formula to calculate spatial proximity using latitude and longitude coordinates relative to the user's latest scheduled daily activity.

Multi-Tenant Collaboration & Security Matrix

• Challenge: Decoupling public itinerary visibility from private collaborative editing privileges without introducing security vulnerabilities or permission leaks.

• Solution: Designed a relational permission schema separating global access types from specific user collaboration roles. This allowed owners to maintain public/private visibility states while delegating selective read or write access to individual collaborators.

Transactional Integrity & API Fault Tolerance

• Challenge: Preventing partial data corruption during multi-record operations like itinerary cloning, while maintaining recommendation availability during third-party API downtime.

• Solution: Wrapped multi-record database operations in explicit MySQL database transactions to ensure strict atomicity, while implementing an offline local fallback strategy when Google Places API calls fail or hit rate limits.