As an extension of our study "High-Speed Railways and the Geography of Inventors' Collaboration: Evidence from France (1980-2010)", this paper provides a detailed account of our methodology for constructing a novel dataset of city-to-city train travel times in France from 1980-2020. We use an arrival-departure time schedule from Société Nationale des Chemins de Fer (SCNF), the French national state-owned railways company, as well as dates of high-speed railways (HSR) openings. From 1981 to 2017, high-speed lines have been built to connect Paris to major cities in France. Using Dijkstra algorithm, we compute the contemporaneous travel time between every two cities in France. Then, to compute the past values of travel time within each pair, we rely on the assumption that prior to an HSR opening, trains were running at a normal speed. We are able to compare our estimations of travel time by train to observed values of travel time from a subsample of city-pairs (SNCF). Our dataset is found to replicate 95% of the variation in the observed travel time.