ASF² Initiative
ASF² is building a new generation of spatial, theory-driven geospatial intelligence.
If you care about spatial structure, interpretability, and methodological depth, not just model accuracy, we invite you to join us.
Re-thinking spatial features as systems
The Advanced Spatial Feature Systems (ASF²) Initiative is an international, theory-driven research initiative dedicated to advancing the foundations of geospatial analysis.
ASFS focuses on the development of spatial feature systems, moving beyond ad-hoc feature engineering toward principled, interpretable, and scalable spatial theories and methods.
At the core of the initiative is the idea of ASF² — Spatial Features → Spatial Systems.
Rather than treating spatial features as isolated inputs to models, we seek to formalise how spatial features interact, organise, and scale as coherent systems that reflect spatial structure, heterogeneity, and dependence.
We are building a new generation of geospatial theory and methods that place spatial structure first, not as an afterthought to model performance.
What is ASF²?
ASF² represents a fundamental shift in geospatial modelling. While most contemporary workflows prioritise predictive accuracy through increasingly complex algorithms, ASF² redirects attention to the foundations of spatial modelling by asking what constitutes a valid spatial feature and how such features should be constructed, interpreted, and evaluated.
Rather than treating spatial features as isolated model inputs, ASF² views them as interacting systems that operate across space and scale. It explicitly addresses how spatial structure, heterogeneity, and local anomalies can be encoded in a principled manner and how these properties shape spatial dependence and spatial behaviour.
By adopting this perspective, ASF² provides a unifying lens for developing spatial feature hierarchies, system-level spatial indicators, theory-grounded validation frameworks, and interpretable spatial learning methods, enabling spatial structure to be treated as a first-class component of geospatial analysis rather than an afterthought to model performance.
Why Join ASF²?
Joining ASF² means contributing to foundational research that shapes how spatial data are modelled, interpreted, and validated.
ASF² offers opportunities to develop new spatial theories, indicators, and methods; contribute to flagship methodological papers and frameworks; collaborate in an international and intellectually rigorous research environment; and build a strong theoretical foundation that transfers across application domains while establishing a clear academic identity in spatial methods and GIScience.
Mentorship within ASF² prioritises intellectual ownership and transparent authorship, critical thinking over routine coding, and long-term research vision.
Who We Are Looking For
We seek students and early-career researchers who are committed to advancing spatial theory and methods, rather than merely applying black-box models.
Suitable applicants are those with a strong interest in spatial statistics, GIScience, or geospatial theory; who care about why methods work, not only whether they improve accuracy; who are comfortable engaging with definitions, assumptions, and validation logic; and who enjoy method-driven, conceptually rigorous research.
Relevant backgrounds include spatial statistics and geostatistics, GIScience and spatial analysis, GeoAI and spatial machine learning, remote sensing and Earth system science, and applied mathematics or data science with a spatial focus. Experience in R and/or Python, working with spatial data and spatial dependence, familiarity with heterogeneity, scale effects, or spatial diagnostics, and experience with methodological research are advantageous but not mandatory.
ASF² is not suitable for applicants seeking application-only projects or purely black-box modelling work.
Reference
Spatial association
Song, Y., 2022. The second dimension of spatial association. International Journal of Applied Earth Observation and Geoinformation. 111, 102834. pdf
Spatial validation
Liu, H., Song, Y. and Yi, W., 2026. Degree of spatial interpretability. International Journal of Geographical Information Science, pp.1-21. pdf
Geocomplexity
Zhang. Z., Song, Y., Luo, P., & Wu, P., 2023. Geocomplexity explains spatial errors. International Journal of Geographical Information Science. pdf
Spatial heterogeneous features
Zhang, Z., Li, Z. and Song, Y., 2024. On ignoring the heterogeneity in spatial autocorrelation: consequences and solutions. International Journal of Geographical Information Science, pp.1-27. pdf
Yang, X., Song, Y., Yoo, C., Ren, K. and Wu, P., 2025. Irregular anisotropy in surface urban heat island footprint. Sustainable Cities and Society, p.106779. pdf
Spatial outliers
Ren, K., Song, Y. and Yu, Q., 2025. Second-dimension outliers for spatial prediction. International Journal of Geographical Information Science. pdf
Local power
Hu, J., Song, Y. and Zhang, T., 2025. A local indicator of stratified power. International Journal of Geographical Information Science. pdf
Sun, Y., Wu, G., Song, Y., Liu, H., Wang, L., Zhang, Z. & Hu, J., 2026. Local effects of pattern interactions in driving urbanization. International Journal of Applied Earth Observation and Geoinformation, 146, 105072. pdf
Hu, J., Qu, R., Song, Y. and Wu, P., 2025. Local pathways of association. International Journal of Applied Earth Observation and Geoinformation, 139, p.104531. pdf
Spatial heterogeneity
Luo., P., Song, Y., Zhu, D., Cheng, J., & Meng, L., 2023. A generalized heterogeneity model for spatial interpolation. International Journal of Geographical Information Science. 37(3): 634-659. pdf
Luo, P., Li, Y., Song, Y., Li, Z. and Meng, L., 2025. Measuring univariate effects in the interaction of geographical patterns. International Journal of Geographical Information Science, pp.1-32. pdf
Zhang, Z., Song, Y., Karunaratne, L. and Wu, P., 2024. Robust interaction detector: A case of road life expectancy analysis. Spatial Statistics, p.100814. pdf
Luo, P., Song, Y., Huang, X., Ma, H., et al., 2022. Identifying determinants of spatio-temporal disparities in soil moisture of the Northern Hemisphere using a geographically optimal zones-based heterogeneity model. ISPRS Journal of Photogrammetry and Remote Sensing. 185, 111-128. pdf
Zhang, Z., Song, Y., & Wu, P., 2022. Robust geographical detector. International Journal of Applied Earth Observation and Geoinformation. 109, 102782. pdf
Song, Y., Wang, J.F., Ge, Y., et al., 2020. An optimal parameters-based geographical detector model enhances geographic characteristics of explanatory variables for spatial heterogeneity analysis: cases with different types of spatial data. GIScience & Remote Sensing. 57(5): 593-610. pdf
Spatial Methods 