Real-Time Logistics & Fleet Management

The Challenge

Blind Spots Across the Entire Fleet

Our client operated a fleet of over 500 delivery vehicles across British Columbia, serving everything from same-day e-commerce parcels to scheduled B2B shipments. Despite the scale of their operation, dispatchers were still planning routes using spreadsheets and static maps. There was no real-time visibility into where vehicles were, whether drivers were on schedule, or if routes were being followed.

Fuel costs had become the single largest line item on their P&L, and the manual routing process meant that drivers were frequently taking suboptimal paths, idling in traffic, or making unnecessary detours. Customer complaints about late deliveries were increasing month over month, and the on-time delivery rate had fallen below 80%.

The client needed a unified platform that could track every vehicle in real time, automatically optimize routes based on live traffic and delivery constraints, and give operations managers a single pane of glass to monitor performance, identify bottlenecks, and make data-driven decisions.

Our Solution

IoT, Machine Learning, and Real-Time Visibility

We started by integrating IoT GPS trackers across the entire fleet using AWS IoT Core as the ingestion layer. Each vehicle reported its position, speed, fuel level, and engine diagnostics every 10 seconds via MQTT. This telemetry data flowed into a time-series database, giving us a rich historical dataset for analysis and a real-time stream for live tracking.

The route optimization engine was built with Python and TensorFlow. We trained a model on 18 months of historical delivery data combined with real-time traffic feeds, weather conditions, and delivery time windows. The model generates optimized routes each morning and dynamically re-routes vehicles throughout the day as conditions change — accounting for traffic incidents, new priority orders, and driver break schedules.

The operations dashboard was built with React and MapBox GL, rendering real-time vehicle positions on an interactive map with color-coded status indicators, geofence alerts, and delivery progress tracking. Dispatchers could click any vehicle to see its current route, estimated arrival times, and historical performance metrics. PostgreSQL powered the analytics backend, supporting complex queries across millions of delivery records for reporting and forecasting.

“Before Mapletech Labs, we were flying blind. Now we have real-time visibility into every vehicle, every route, every delivery. The fuel savings alone paid for the entire project within six months. Our dispatchers went from managing chaos to managing optimization — and our customers have noticed the difference.”