Title

AWS re:Invent 2022 - [NEW] Easily build, train, and deploy ML models using geospatial data (AIM218)

Summary

The session is aimed at data scientists, machine learning engineers, product managers, geospatial analysts, and executives interested in geospatial ML.
Geospatial data includes aerial and satellite imagery and map data, which has become more accessible due to cost reductions and the prevalence of location data emitting devices.
Use cases for geospatial data in ML span across insurance, commodity trading, government policy, city planning, agriculture, and retail.
Despite the availability of geospatial data, its use in ML is limited due to the complexity of accessing and processing the data, and the specialized skills required.
AWS introduces new geospatial capabilities in SageMaker to address these pain points, including pre-built integrations for open data sets, geospatial operations, pre-installed geospatial libraries, pre-trained models, and interactive maps for visualization.
A demo showcases how SageMaker's geospatial capabilities can be used to monitor the shrinking of Lake Mead.
Damien from Arup presents a case study on using SageMaker geospatial capabilities for urban planning and understanding the urban heat island effect.

The reduction in satellite operation costs and the ubiquity of location data emitting devices have democratized access to geospatial data, enabling new applications and insights.
The complexity of geospatial data workflows has been a significant barrier to the wider adoption of geospatial ML, with challenges in data access, processing, and visualization.
AWS SageMaker's new geospatial capabilities aim to streamline the geospatial ML workflow, making it more accessible and reducing the time from data access to inference.
Pre-built integrations and pre-trained models can significantly reduce the time and expertise required to start working with geospatial data in ML applications.
The interactive map feature within SageMaker can facilitate collaboration between data scientists and other team members, improving organizational efficiency.
The case studies presented demonstrate the practical impact of AWS SageMaker's geospatial capabilities on industries such as agriculture and urban planning, highlighting the potential for improved decision-making and efficiency gains.