Modeling the Determinants of Urban Short-Term Rental Prices in Gambia: A Case of Greater Banjul Area

Short-term rentals (STR) are forming a growing part of the accommodation offer to international and domestic tourists. Yet, the drivers of their prices and spatial structures are understood poorly in sub-Saharan African settings. The present study explores the determinants of STR pricing in the GBA, focusing on spatial proximity to POIs, attributes of accommodation, and hosts. Based on an enriched set of 5657 Airbnb listings (September 2024–July 2025), we performed spatio-temporal feature engineering including booking seasonality, length of stay, and distances to beaches, tourist, and entertainment areas. The price distribution showed that the market was mainly budget to midrange (USD 20–60 per night), with a few high-end accommodation options. Correlation and model interpretation analyses showed that coastal proximity, property size, and entire-home listings significantly increase prices, while distance from amenities and non-super-host status reduce them. Based on the six machine learning prediction models, Random Forest and XGBoost had the highest prediction accuracy (R² = 0.92–0.93). The results from SHAP analysis confirm that the determinant of price in GBA mainly depends on location features. The results inform spatially targeted policy strategies of tourism development, affordability, and sustainable urban development, and provide new empirical evidence on STR urban patterns in Africa.