Machine Learning Platform

AI-powered modelling and analytics tools

We developed the ORR Model Builder, an advanced machine learning and scenario simulation platform built with Python and Node.js, enabling the Office of Rail and Road to explore complex operational questions without requiring in-house coding expertise. The platform empowers ORR teams to design, train, and deploy predictive models that identify patterns in operational data, simulate the impact of different policy or scheduling decisions, and forecast future network performance under varying conditions.

Applications include rail timetable optimisation, capacity planning, maintenance scheduling, and incident impact forecasting, allowing planners to test multiple “what-if” scenarios before committing to real-world changes. The platform integrates automated data ingestion, feature engineering, and model selection, ensuring accuracy and repeatability.

The cloud-based GUI, hosted in ORR’s Microsoft Azure environment, offers interactive visual analytics, drag-and-drop model configuration, and real-time scenario playback, providing decision-makers with a dynamic, data-driven environment to evaluate operational strategies.

Alongside this, we built the Daily Logs Explorer, a powerful NLP-driven text analytics tool for mining ORR’s vast archives of operational reports and logs. Using entity recognition and sentiment analysis, it surfaces incident patterns, highlights safety risks, and links historical events to potential future impacts — creating a richer evidence base for scenario modelling and strategic decision-making.

‍