Delay patterns in UK rail (dissertation)

wip · 2026-03-22

MSc Data Science dissertation on temporal patterns of delays, cancellations and service reliability in UK rail. In progress

Background

What passengers actually feel is not the average delay. It is the reliability of their hour and their operator. The topic: what patterns do UK rail delays and cancellations show across time of day, period and train operating company, and how far can public data alone take that question?

Design

Built on ORR public statistics, layered in order: descriptive statistics, temporal trend analysis, cross-operator comparison, then an interactive dashboard. All in R. Deliberately no causal inference or forecasting from the start. First pin down the questions public data can answer honestly (describing and comparing patterns), then decide whether supplementary data like weather earns its place on top.

Outcome

In progress. Target outputs: analysis code in R, a reproducible data pipeline, and a dashboard for exploring reliability by operator and time. The plan is to put that dashboard on this site’s lab subdomain when it is done.