From Time and Place to Preference: LLM-Driven Geo-Temporal Context in Recommendations

TL;DR

This paper introduces a scalable LLM-based framework for generating geo-temporal embeddings from minimal data, capturing contextual factors like holidays and events to improve recommendation systems.

Contribution

It presents a novel method to create geo-temporal embeddings using LLMs and demonstrates their effectiveness in enhancing recommendation models.

Findings

Geo-temporal embeddings provide predictive signals aligned with full model outcomes.

Incorporating embeddings improves recommendation accuracy.

The framework is validated on MovieLens, LastFM, and a production dataset.

Abstract

Most recommender systems treat timestamps as numeric or cyclical values, overlooking real-world context such as holidays, events, and seasonal patterns. We propose a scalable framework that uses large language models (LLMs) to generate geo-temporal embeddings from only a timestamp and coarse location, capturing holidays, seasonal trends, and local/global events. We then introduce a geo-temporal embedding informativeness test as a lightweight diagnostic, demonstrating on MovieLens, LastFM, and a production dataset that these embeddings provide predictive signal consistent with the outcomes of full model integrations. Geo-temporal embeddings are incorporated into sequential models through (1) direct feature fusion with metadata embeddings or (2) an auxiliary loss that enforces semantic and geo-temporal alignment. Our findings highlight the need for adaptive or hybrid recommendation strategies, and we release a context-enriched MovieLens dataset to support future research.