Condense Powered Airport & Hotel Shuttle Optimization

Business Context 

Airport-to-hotel shuttle services represent a critical touchpoint in guest experience for travel and hospitality operators. These operations, often managed with static schedules and manual coordination, are prone to inefficiencies such as vehicle idling, guest wait times, and missed pickups. Delays in flight arrivals, real-time changes in guest behavior, and urban traffic dynamics further complicate these systems. 

This document outlines how Condense, a vertically optimized real-time streaming platform, enables organizations to design and deploy a fully automated, event-driven shuttle dispatch system. By unifying real-time data sources such as flight schedules, hotel reservations, live traffic, and shuttle telemetry, Condense empowers hospitality providers to deliver faster, more intelligent, and more efficient guest transport experiences. 

Objectives 

• Establish a real-time shuttle dispatch system powered by streaming data. 

• Integrate flight, reservation, traffic, and fleet telemetry into a unified pipeline. 

• Automate routing, vehicle assignment, and guest communication. 

• Maintain high availability, operational visibility, and scalability. 

• Minimize guest wait time and optimize fleet usage across properties. 

Problem Statement 

Current Challenges 

• Disjointed Data Systems: Lack of real-time coordination between flight arrivals and hotel bookings. 

• Manual Dispatching: Front desk or call-based dispatching introduces delays and human error. 

• Fleet Underutilization: Vehicles operate below capacity or remain idle due to poor scheduling. 

• Limited Visibility: Operators lack a unified view of guest location, shuttle readiness, and road status. 

• Scalability Bottlenecks: Fragmented integration approaches do not scale across multiple hotels or hubs. 

Technical Requirements 

To resolve these challenges, the system must be built on: 

• Event-Driven Architecture: React to real-time changes in flights, bookings, and traffic. 

• Stream Processing Engine: Enrich, correlate, and transform event data on the fly. 

• State-Aware Computation: Track guest readiness, vehicle availability, and group assignments. 

• Integrated Action Layer: Automatically trigger dispatch actions, app updates, and guest notifications. 

Data Ingestion & Schema Management 

Condense provides prebuilt, schema-aware connectors tailored for travel and hospitality workflows. These connectors support real-time ingestion, schema enforcement, and transformation at scale. 

A. Flight Status Connector 

• Type: REST API (poller or webhook) 

• Prebuilt Support: AviationStack, FlightAware 

• Fields: flight_id, arrival_time, status, gate, updated_at 

• Deployment: Configured as a connector block in the Condense pipeline with automatic schema binding 

B. Hotel Reservation Connector 

• Type: REST-based integration with PMS 

• Mapping: Condense supports mapping guest metadata with schema (e.g., flight_id, check_in) 

• Stream Conversion: Transforms batch API responses into live streams (hotel.reservation) 

C. Traffic ETA Connector 

• Source: Google Maps / TomTom API / MapmyIndia 

• Polling Frequency: Every 2–5 minutes 

• Result Fields: origin, destination, eta_seconds, traffic_level 

• Transform Options: Delay classification using built-in conditional transform block 

D. Fleet GPS Connector 

• Protocol: MQTT / Kafka / HTTP 

• Fields: vehicle_id, lat_lng, occupancy, status 

• Output Stream: fleet.location 

Data Processing 

Condense enables powerful real-time processing through both no-code logic blocks (NCLC) and a developer IDE supporting polyglot programming. 

A. Matching Guests with Flights 

• Merge Utility: Joins flight.status and hotel.reservation streams on flight_id 

• Window Utility: Applies a 3-hour time window to accommodate early/late arrivals 

• Result: An enriched guest.arrival stream with real-time flight alignment 

B. Traffic-Based Scheduling 

• Window Utility: Monitors changing route.eta to detect delays 

• Alert Utility: Flags events when ETA exceeds historical thresholds 

• Transform Output: Adjusts planned pickup time or assigns an alternate vehicle 

C. Fleet Availability Evaluation 

• Split Utility: Separates shuttles by availability and proximity 

• Alert Utility: Notifies if a vehicle is idle for an extended duration 

• State Store: Maintains real-time vehicle status and previous assignments 

Custom Transform via Condense Inbuilt IDE 

For advanced routing, optimization, or business rule logic, Condense provides an inbuilt IDE with Git integration and multi-language support. 

Developer Workflow 

Author custom logic in Python, JavaScript, Java, Go, or any supported language Edit, test, and validate directly inside the Condense IDE Integrate with Git repositories for version control, rollback, and CI/CD Deploy transforms live on production pipelines with real-time data validation 

Example Use Cases:  

• Proximity-based shuttle assignment 

• Grouping passengers by destination, arrival time, or VIP status 

• Predictive ETA modeling using rolling averages or traffic heuristics 

• Enriched guest notifications with dynamic ETA, driver info, and language preference 

Sample Triggering Conditions 

• A shuttle is available 

• Guest group ETA is within a predefined time window 

• Traffic delay is under a configured threshold 

Output Actions

• Send dispatch command via REST API to the driver app (/assign_shuttle) 

• Push to the guest.notifications.sms queue for outbound communication 

• Update operational dashboard with dispatch.status 

Integration Touchpoints and Downstream Connectors 

Condense includes built-in downstream connectors for seamless system interaction. 

A. Kafka Broker Integration 

Used for delivering structured dispatch events (dispatch.command) to enterprise systems. 

B. REST/Webhook Connector 

Supports integration with driver mobile apps, real-time maps, and control panels. 

C. Driver App Interaction 

• API Endpoint: /assign_shuttle 

• Payload: group_id, pickup_location, ETA, guest_count 

• Response Handling: Waits for ACK/NACK; retries up to 3 times before alerting dashboard 

D. Guest Notification Service 

• Channels Supported: SMS (Twilio, Gupshup), WhatsApp, Email 

• Payload: Shuttle details, estimated pickup time, driver contact 

• Trigger Events: Dispatch assignment, route update, or delay notification 

E. Operational Dashboard 

• Consumes dispatch.status, fleet.telemetry, and guest.grouping 

• Displays real-time vehicle tracking, pickup status, and alerts 

F. Storage Sink Integration 

• Persist enriched event streams to S3, Azure Blob, or Google Cloud Storage 

• Supports compliance, historical analysis, and data science use cases 

Observability, Resilience, and Operational Monitoring 

Condense is designed for production-grade reliability with built-in observability tools. 

Metrics Tracked 

Average shuttle wait time per group 

Computation 

• Each guest group receives an estimated pickup time (scheduled_pickup_time) when the dispatch is planned. 

• The actual pickup event (from vehicle GPS stream or driver confirmation) is timestamped as actual_pickup_time. 

• The wait time is computed as the difference between these two timestamps for each group. 

Implementation in Condense: 

• Use a Join transform to correlate guest.grouping and pickup.confirmation streams on group_id. 

• Apply a Custom Transform (code or no-code) to calculate the wait time delta. 

• Output this metric to a shuttle.wait_time.metrics stream for real-time tracking and alerting. 

Visualization can be achieved in PowerBi or any Dashboard tools 

• Dashboard panel showing average wait time per group, updated in real-time. 

• Optionally segmented by hotel or time of day. 

SLA adherence for group pickups 

Computation 

• The SLA for pickup (e.g., within 10 minutes of scheduled time) is defined as a configurable threshold. 

• Wait times exceeding the threshold are marked as SLA violations. 

Implementation in Condense

• Use the Conditional Filter utility to classify each group as SLA_met or SLA_violated. 

• Track total and violating counts over a rolling time window using Windowed Aggregation. 

• Output to sla.adherence.metrics stream. 

Visualization can be achieved in PowerBi or any Dashboard tools 

• SLA compliance percentage shown on the dashboard (e.g., 92.5% SLA adherence today). 

• Alerts can be configured if SLA adherence drops below a defined benchmark. 

Vehicle idle time by hour 

Computation 

• Each vehicle’s state (e.g., active, idle, assigned, in_transit) is tracked via the fleet.location stream. 

• When a vehicle enters idle state, a timer starts. When it becomes active again, the idle duration is computed. 

Implementation in Condense: 

• Use a State Store to maintain the last active state timestamp for each vehicle_id. 

• A Windowed Join or Duration Calculator measures how long a vehicle remained idle. 

• Output to fleet.idle_time.metrics stream. 

Visualization can be achieved in PowerBi or any Dashboard tools 

• Per-vehicle or per-depot idle time breakdown by hour or shift. 

• Alerts on vehicles idle for more than a configured duration (e.g., 30 minutes). 

Dispatch command success rate 

Computation 

• Every dispatch command issued (dispatch.command) expects a response from the driver or system (dispatch.acknowledgement). 

• Success is determined by a valid acknowledgment within a time window (e.g., 60 seconds). 

  
  

Implementation in Condense: 

• Join the dispatch.command and dispatch.acknowledgement streams by dispatch_id. 

• Classify each as success, retry, or failure based on timing and acknowledgment status. 

• Aggregate counts using a Windowed Counter and calculate success rate over time. 

Visualization can be achieved in PowerBi or any Dashboard tools 

• Live chart of success rate (e.g., 98% success in last 15 minutes). 

• Drill-down view by vehicle or driver to identify systemic issues. 

Failure Scenarios and Mitigations 

Observability Features 

• Live Stream View: Inspect input and output of each transform in real-time 

• Execution Logs: Debug individual record flows across pipeline stages 

• Dashboard Builder: Create real-time dashboards directly within Condense 

• Alert Builder: Configure automated alerts to Slack, Email, or SMS channels 

Retry and Dead Letter Handling 

• Retries configurable per connector or transform 

• Failed events routed to DLQ with root-cause tagging 

• Manual replay and correction supported via platform interface 

Platform Advantage – Why Condense is Uniquely Positioned 

Condense is not a general-purpose streaming engine—it is a purpose-built, vertically optimized real-time application platform. Its core advantages in this use case include: 

Industry-Specific Modules 

Condense provides prebuilt connectors and transforms tailored for travel and hospitality domains, eliminating the need for custom data ingestion or basic transformation code. This drastically accelerates development. 

Developer + No-Code Synergy 

Condense enables business teams to use NCLC logic utilities while giving developers complete control through inbuilt IDE. Complex, stateful logic can be authored in Python, JavaScript, Java, or any language of choice. 

End-to-End Real-Time Pipeline 

The platform provides an integrated experience—from data ingestion and enrichment to dispatch decisioning and alerting—without needing external orchestration tools, schedulers, or microservices. 

GitOps and Production Readiness 

Integrated Git support allows full version control, deployment traceability, and CI/CD integration, enabling enterprise-grade governance and reproducibility. 

Continuous Data Workflows, Uninterrupted 

Condense ensures uninterrupted execution of transforms for continuous streaming data processing, delivering precise outputs to end applications. From hosting to running your applications, Condense manages the entire backend with guaranteed uptime and effortless scalability. Whether scaling during peak loads or maintaining stability for critical workflows, our fully managed platform adapts to your demands with ease. With consistent performance and reliability, Condense empowers your business to thrive every step of the way. 

Summary and Strategic Fit 

The Airport and Hotel Shuttle Optimization use case demands a robust, scalable, and intelligent event-driven system. The Condense platform fulfills all requirements with minimal engineering overhead, architectural simplicity, maximum adaptability and deployment efficiency. 

Condense enables this system to be designed, deployed, and scaled entirely within its platform, transforming streaming data into operational automation in record time. 

Condense provides a complete framework for building domain-specific real-time applications. It abstracts away the complexities of stream management while offering full control over transformation, enrichment, and decision logic. 

This shuttle optimization use case is a prime example of how Condense delivers strategic value: 

• Rapid implementation through domain-ready modules 

• Flexible developer tooling for real-time logic deployment 

• Scalable, observable, production-ready pipelines 

• Seamless integration with enterprise systems and end-user applications 

Condense platform is ideal for enterprises that value short implementation cycles, high system resilience, and future-proofed architecture. Condense enables organizations to convert streaming data into real-world actions with speed, scale, and precision.