Special session 5:
Putting Information into Spatial Structures: Approaches in Localising Information and Knowledge

Advances in computational methods have enabled researchers to process and analyse large datasets, but synthesising unstructured information into spatial frameworks remains a challenge. Meaningful information, such as the content of policy documents, media articles or academic research outputs are often only implicitly spatial, are not available on a desired spatial scale or are lacking geographic references completely. Methods such as image recognition, spatial dis-aggregation, geocoding with LLM (Large Language Models) or other machine learning approaches can help derive geographic context from unstructured or coarse data. However, challenges remain in resolving ambiguity, connecting heterogeneous datasets, and accurately interpreting geographic settings.

​

We invite scholars to share their experiences in spatial data synthesis and discuss how these methods enhance our understanding of spatial relationships and geographic patterns. Specifically, we welcome contributions that address topics such as:

- Geoparsing: linking unstructured text documents to spatial locations or objects
- Data synthesis: integrating datasets from various sources to generate new insights
- Social media analysis: localising content from discussions on social media platforms
- Data disaggregation: improving the granularity of spatial datasets
- Development and application of LLMs: using topic modelling for specific thematic areas (e.g., sustainability)
- Synthesising large datasets: transforming vast amounts of information into meaningful knowledge assets
- Enhancing dataset quality: improving the accuracy and reliability of spatial datasets
- Big data strategies: leveraging data analysis techniques to support sustainable urban and spatial development
- Pattern analysis: uncovering insights from text data and large unstructured datasets

​

We also welcome contributions that explore related challenges and innovative approaches beyond these topics.