Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Unlocking Location Intelligence: A Survey from Deep Learning to The LLM Era

Xixuan Hao, Yutian Jiang, Xingchen Zou, Jiabo Liu, Yifang Yin, Yuxuan Liang

TL;DR

This survey maps the evolution of geospatial representation learning from deep learning-driven, modality-specific methods to cross-modal and LLM-enabled approaches in the Location Intelligence (LI) domain. It introduces a three-dimensional taxonomy—data perspective, methodology perspective, and application perspective—and partitions methods into single-view, dual-view, and multi-view embeddings for both locations and regions. The study synthesizes current advances, identifies limitations such as benchmarking gaps and foundation-model integration, and outlines future directions toward unified benchmarks, large-scale multimodal datasets, and LLM-driven geospatial reasoning. The findings underscore the potential of multimodal GeoAI to enhance socioeconomic prediction, regional planning, and location-aware decision support in urban environments.

Abstract

Location Intelligence (LI), the science of transforming location-centric geospatial data into actionable knowledge, has become a cornerstone of modern spatial decision-making. The rapid evolution of Geospatial Representation Learning is fundamentally reshaping LI development through two successive technological revolutions: the deep learning breakthrough and the emerging large language model (LLM) paradigm. While deep neural networks (DNNs) have demonstrated remarkable success in automated feature extraction from structured geospatial data (e.g., satellite imagery, GPS trajectories), the recent integration of LLMs introduces transformative capabilities for cross-modal geospatial reasoning and unstructured geo-textual data processing. This survey presents a comprehensive review of geospatial representation learning across both technological eras, organizing them into a structured taxonomy based on the complete pipeline comprising: (1) data perspective, (2) methodological perspective and (3) application perspective. We also highlight current advancements, discuss existing limitations, and propose potential future research directions in the LLM era. This work offers a thorough exploration of the field and providing a roadmap for further innovation in LI. The summary of the up-to-date paper list can be found in https://github.com/CityMind-Lab/Awesome-Location-Intelligence and will undergo continuous updates.

Unlocking Location Intelligence: A Survey from Deep Learning to The LLM Era

TL;DR

This survey maps the evolution of geospatial representation learning from deep learning-driven, modality-specific methods to cross-modal and LLM-enabled approaches in the Location Intelligence (LI) domain. It introduces a three-dimensional taxonomy—data perspective, methodology perspective, and application perspective—and partitions methods into single-view, dual-view, and multi-view embeddings for both locations and regions. The study synthesizes current advances, identifies limitations such as benchmarking gaps and foundation-model integration, and outlines future directions toward unified benchmarks, large-scale multimodal datasets, and LLM-driven geospatial reasoning. The findings underscore the potential of multimodal GeoAI to enhance socioeconomic prediction, regional planning, and location-aware decision support in urban environments.

Abstract

Location Intelligence (LI), the science of transforming location-centric geospatial data into actionable knowledge, has become a cornerstone of modern spatial decision-making. The rapid evolution of Geospatial Representation Learning is fundamentally reshaping LI development through two successive technological revolutions: the deep learning breakthrough and the emerging large language model (LLM) paradigm. While deep neural networks (DNNs) have demonstrated remarkable success in automated feature extraction from structured geospatial data (e.g., satellite imagery, GPS trajectories), the recent integration of LLMs introduces transformative capabilities for cross-modal geospatial reasoning and unstructured geo-textual data processing. This survey presents a comprehensive review of geospatial representation learning across both technological eras, organizing them into a structured taxonomy based on the complete pipeline comprising: (1) data perspective, (2) methodological perspective and (3) application perspective. We also highlight current advancements, discuss existing limitations, and propose potential future research directions in the LLM era. This work offers a thorough exploration of the field and providing a roadmap for further innovation in LI. The summary of the up-to-date paper list can be found in https://github.com/CityMind-Lab/Awesome-Location-Intelligence and will undergo continuous updates.
Paper Structure (32 sections, 3 equations, 7 figures, 2 tables)

This paper contains 32 sections, 3 equations, 7 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Preliminaries
  3. Region Representation Learning
  4. Location Representation Learning
  5. Data Modality Perspective
  6. Spatial Data
  7. Mobility Data
  8. Social Media Data
  9. Socio-Economic Attributes
  10. Overall
  11. Methodology Perspective
  12. Data-centric View
  13. Single View
  14. Dual View
  15. Multiple View
  16. ...and 17 more sections

Figures (7)

  • Figure 1: The complete pipeline of geospatial representation learning with location-centric data.
  • Figure 2: The concepts of location, location embedding, region, and region embedding in geospatial representation.
  • Figure 3: The usage frequency of data modalities during learning stage across four categories in the survey. Each category contains commonly used data modalities.
  • Figure 5: A taxonomy of representative works for geospatial representation learning.
  • Figure 6: Taxonomy of Application for Geospatial Embedding.
  • ...and 2 more figures