What is it about?
Geospatial Machine Learning is a growing domain of research that combines geospatial data and machine learning to draw insights from the surrounding world. The publication introduces a new, fully open-source Python package called Spatial Representations for AI (SRAI in short). The library simplifies access to open-source geospatial data and integrates many geo-related algorithms with a unified API. It includes tools for downloading geospatial data split a given area into micro-regions using multiple algorithms and train an embedding model using various architectures Our main goal in developing such a library is to make geospatial data processing, especially Geospatial Machine Learning, easier and more accessible for machine learning practitioners and GIS experts. We believe that by making it easy to use and share geospatial data and models, we can encourage more people to utilize geospatial data in their solutions and improve the maturity of the GeoAI domain.
Featured Image
Photo by NASA on Unsplash
Read the Original
This page is a summary of: SRAI, November 2023, ACM (Association for Computing Machinery),
DOI: 10.1145/3615886.3627740.
You can read the full text:
Resources
SRAI Github Repository
The code repository of the SRAI library.
Kraina.ai research lab webpage
The webpage of the research lab at Wrocław University of Science and Technology responsible for creating SRAI.
Tutorial on Geospatial Machine Learning with SRAI @ EuroScipy 2023 in Basel
A recording of a tutorial on Geospatial Machine Learning with SRAI, we conducted at EuroScipy 2023 in Basel.
Code for Tutorial @ Euroscipy 2023
Repository containing the code for our tutorials. This points to a release for EuroScipy 2023.
Code for Tutorial @ MLinPL 2023
Repository with code accompanying our tutorial at MLinPL 2023 in Warsaw
Contributors
The following have contributed to this page