Jiang, ZhiqiuMondschein, Andrew2020-05-012020-05-012019https://hdl.handle.net/11299/212972Urban planners often seek to establish land use patterns around transit stations that encourage non-auto travel. However, the willingness of travelers to use different modes in the vicinity of transit remains understudied, in part because of the lack of spatially-precise data on destination and mode choices. Using transportation content extracted from Yelp, a location-based social network (LBSN), we investigate how travel mode to non-work destinations is influenced by proximity to transit. We use textual analysis to analyze travel for non-work activities in seven cities across North America and Europe. Mixed-effect and binomial logistic models show how reported travel by mode varies by distance to rail transit stations. We find that for most non-work activity purposes, reported rail use is highly sensitive to proximity to stations, but some purposes are more amenable to rail use overall. Additionally, compared to non-US cities, US cities are far more parking-dependent near rail stations. The results suggest that not all activities elicit the same levels of non-auto travel, and transit-oriented planning should account for specific activities and regional factors that may modify willingness to travel by different modes. While subject to limitations, LBSNs can illuminate local travel with greater spatial specificity than traditional surveys.enRail transitSocial mediaAccessibilityTransit oriented developmentArticle10.5198/jtlu.2019.1409