This is a private repository for peer review. It will be made publicly available upon publication. What's included The prepared analysis dataset is included, so the results and figures can be reproduced without downloading anything. A README with step-by-step instructions, software requirements, and links to the original public data sources. Ten R scripts that run the full analysis in numbered order, from raw-data processing through the final figures. A single configuration file that sets all options and file locations. fdataclean. rds: a panel of daily ridership by docking station, with smoke, weather, and prior-exposure measures already merged, so the analysis and most figures can be reproduced directly without rebuilding it from the raw trip records. The station-location file used by the map figures. The derived station-level daily climate and smoke histories used to assemble the panel. rawclimate. zip: the wildfire-smoke grid, smoke plumes, station distance matrices, and station metadata. How to use it Download the files into a single folder and unzip rawclimate. zip in place. Install R (version 4. 4 or later) and the packages listed in the README, then run the scripts from that folder. Paths resolve automatically, so no editing is required. Because the prepared dataset is included, reviewers can start partway through (script 4) and regenerate the models and figures directly. Data sources The raw bikeshare trip files come from each city's public open-data portal; wildfire-smoke estimates are from Childs et al. (2022) ; climate data are from the PRISM Climate Group. All are publicly available, and the README lists the original access links. The raw trip records are not bundled here, as they are large and freely available from the city portals; they are needed only to rebuild the panel from scratch. The prepared panel and all derived inputs are included, so the analysis and figures reproduce without any downloads.
Carlos Gould (Wed,) studied this question.