Publication Details
Abstract
This study examines the spatial distribution, attraction and satisfaction signals of hotels on Ctrip in Tianjin, China, based on the data collected on August 15, 2026. The original data set contains 993 hotel listings and 155,258 review entries; the geographical coordinates from the hotel information pages were converted from BD-09 to WGS84 and summarized into 983 unique positions. The average nearest-neighbour index of 0.318 (z = -40.89) indicates a high degree of clustering. Kernel density analysis reveals a peak in the central city near 117.195°E, 39.115°N and another peak in the Binhai New Area near 117.687°E, 39.018°N. Hotel scores show weak but significant spatial autocorrelation (Moran’s I = 0.070), whereas the total number of cumulative reviews exhibits a stronger spatial pattern (I = 0.329). An exploratory two-stage least-squares model with 696 complete cases indicates a positive effect of hotel scores (coefficient = 0.201) and a high spatial lag of platform attention (0.745). Reviews about family and business travel account for 52.3% of all comments; however, defects like maintenance and noise are commonly found in the less satisfactory reviews. The hotel market in Tianjin is polycentric: the supply is mainly concentrated in the downtown area, Binhai and Jizhou, whereas the online visibility is more geographically focused than the actual quality perceived. Taking into account the restriction of data acquisition and the spatial residuals, these findings are considered as tentative associations rather than certain causal relationships.