Location Intelligence (LI) focuses on turning geospatial data into actionable knowledge and is crucial for spatial decision-making.
Geospatial Representation Learning has been transforming LI development, with the recent advancements in deep neural networks (DNNs) and large language models (LLMs).
The integration of LLMs has enhanced capabilities for cross-modal geospatial reasoning and processing unstructured geo-textual data.
This survey provides a detailed overview of geospatial representation learning, categorizing advancements into data, methodological, and application perspectives, while also suggesting future research directions in the LLM era.