A multimodal dataset for POI (Point of Interest) recommendation, based on the Yelp Open Dataset.It includes business metadata, user reviews, business photos, and LLM-generated summaries of reviews and images.This dataset supports downstream tasks like session-based recommendation, multimodal embedding learning, and more.
business.csv
Business metadata including name, address, categories, etc.… See the full description on the dataset page:
https://huggingface.co/datasets/wzehui/Yelp-Multimodal-Recommendation.