HASTODO: Harnessing Multisource Data for Public Transport

Funder: Royal Society | Partners: DTU

HArnessing multiSource data for public TranspOrt: problem Diagnosis and stochastic Optimisation (HASTODO)

🚀 Overview

Despite continuous investment in public transport infrastructure, current systems often struggle to meet traveller expectations and achieve desired market shares. The HASTODO project addresses this gap by postulating that traditional planning—relying heavily on internal data like passenger surveys and smart cards—overlooks the broader mobility ecosystem.

We propose a holistic, data-driven methodology that harnesses heterogeneous multisource data to unlock the full potential of public transportation.

🔍 The Challenge

Prevailing methods for designing public transport services often suffer from a "data myopia":

• Narrow Focus: Reliance on internal data (e.g., smart cards) ignores the needs of potential passengers currently using cars or other modes.
• Hidden Problems: Issues impeding modal shift, such as accessibility gaps or equity concerns, remain undiagnosed.
• Under-Optimisation: Services are optimised for existing users rather than capturing latent demand.

💡 Our Solution

HASTODO aims to revolutionise public transport planning through a three-stage approach:

1. Market Analysis

We will harness multisource data to analyse dynamic spatiotemporal travel patterns. By understanding the characteristics and preferences of all commuters—not just current riders—we can identify where the true market potential lies.

2. Hierarchical Diagnosis

We are developing a diagnosis method to identify specific barriers preventing car commuters from switching to public transport. This analysis operates at two levels:

• Macro Level: Examining network-wide service coverage and connectivity.
• Micro Level: Assessing local accessibility and equity constraints.

3. Stochastic Optimisation

Based on our diagnosis, we will devise a multi-objective stochastic optimisation model for network design and timetabling. This model will:

• Improve Key Metrics: Focus on enhancing service coverage, equity, and accessibility to increase market share.
• Handle Uncertainty: Calibrate density functions to characterise various sources of uncertainty, using them as constraints and for scenario generation.
• Leverage AI & OR: Employ state-of-the-art Artificial Intelligence, Operations Research, and Machine Learning techniques to solve complex network design problems efficiently.

🌍 Impact