Towards data-driven policy development: the case of London's built cycling infrastructure

In 2013, £913m of funds was allocated over 10 years for investment in London's cycling infrastructure. Much of this - including guided quietways, protected cycle superhighways and London's crossrail for the bike - opened in summer 2016. The chief objective: to make cycling 'a normal part of everyday life [...] something people hardly think about [...and] something everyone feels comfortable doing' (Greater London Authority 2013).

Traditionally, attempts to evaluate such interventions might rely on survey data describing changes in *claimed* behaviour or high-level data from Automatic Traffic Counters describing infrastructure occupancy. The former are often expensive to collect and suffer from numerous (well-documented) biases and the latter are too high-level to capture more subtle changes in behaviour.

This project will instead use new, large-scale observational datasets - from London's bikeshare, underground and bus network, from route planning services (CycleStreets.net), user-contributed and social media data -- to describe changes in city-wide cycling behaviours pre- and post- the intervention. Crucially, it will identify rich detail around the impact of current investment on behaviour and contribute quantified estimates, under uncertainty, around the impact of future investment.