Bicycling is an increasingly popular mode of travel in Canadian urban areas, like the Greater Toronto and Hamilton Area (GTHA). While trip origins and destinations can be inferred from travel surveys, data on route choice is often not collected which makes it challenging to capture the attributes of routes travelled by people who cycle. With new algorithms for cycle routing it is now possible to infer routes. Using bicycle trip records from the most recent regional travel survey, a spatial interaction model is developed to investigate the built environment correlates of bicycling flows in Hamilton, Ontario, a mid-sized city part of the GTHA. A feature of the analysis is the use of CycleStreets to compare the distance and time according to different routes inferred between trip zones of origin and destination. In addition, network autocorrelation is accounted for in the estimated models. The most parsimonious model suggests that shortest-path quietest routes that minimize traffic best explain the pattern of bicycle trip flows in Hamilton. Commercial and office locations and points of interest at the zone of origin negatively correlate with the production of trips, while different land uses and the availability of jobs at the zone of destination are trip attractors. The use of a route planner offers a novel approach to modelling and understanding bicycling flows within a city. This may be useful for transportation planners to infer different types of routes that bicyclists may seek out and consider these in travel demand models.