OptimalTransit Hub: An End-to-End Analytics and Optimisation Platform

🎯 Project Overview

Author's Note: This interactive dashboard was developed independently by myself. The underlying research is funded by the Royal Society, as part of the HASTODO project in collaboration with Prof. Shaopeng Zhong.

🌐 Live Prototype: Access the Dashboard Here

OptimalTransit Hub is an end-to-end analytics and optimisation platform for urban public transit systems. It fuses multisource transit data — real-time Automatic Vehicle Location (AVL) feeds, General Transit Feed Specification (GTFS) schedules, and operational records — into a unified data pipeline that diagnoses service problems and prescribes evidence-based improvements through stochastic optimisation.

Why This Project?

Urban bus networks face persistent reliability challenges: unpredictable delays, schedule deviations, and poor transfer coordination between routes. Traditional approaches rely on static timetable analysis, missing the dynamic nature of real-world operations. This project bridges that gap by combining real-time data processing with operations research methodologies to deliver actionable insights at both the tactical and strategic levels.

What It Does

1. Data Pipeline — Ingests raw AVL GPS pings and GTFS schedules, sanitises and matches them, then calculates per-trip delay metrics with confidence scoring.
2. Descriptive Analytics — Visualises network coverage, punctuality, headway regularity, CO₂ emissions, and energy consumption across routes and time periods.
3. Diagnostic Analytics — Identifies systemic bottlenecks, root causes of delay (e.g. signal misalignment, congestion corridors), and quantifies their impact through heatmaps and factor analysis.
4. Stochastic Optimisation — Applies Adaptive Large Neighbourhood Search (ALNS) to optimise transfer synchronisation across routes, improving passenger connectivity while respecting operational constraints.
5. Interactive Dashboard — A glassmorphism-themed, mobile-responsive web dashboard that unifies all four analytics layers into a single operational interface.

Key Features

Professional Interface

The platform features a glassmorphism-themed, mobile-responsive web dashboard designed for operational efficiency. By unifying all four analytics layers into a single interface, it provides a seamless transition from data ingestion to actionable prescriptive insights.

• 📖 Built-in technical handbook with pipeline and methodology documentation
• 📱 Mobile-responsive design across phone, tablet, and desktop

This output demonstrates our lab's commitment to translating complex stochastic optimisation research into practical, data-driven tools for sustainable urban mobility.