Using LDA to generalize wordclouds for Airbnb Listing’s features across London. Word2Vec model is used to embed listings features into multi-dimensional vectors and visualize them by UMAP. Comparing embedding vectors to London House Price Index and build prediction model by SVM, which identify and specify listing's homogeneity and heterogeneity more clearly in various regions and boroughs, aiming to help to manage and promote listings better.
Groupwork of CASA0013_FSDS in UCL Master's programme in Urban Spatial Science