GeoAI at RCC GIS

What is GeoAI?

Geospatial Artificial Intelligence (GeoAI) is an interdisciplinary field that combines geospatial science, artificial intelligence (AI), and machine learning (ML) to analyze and interpret spatial data. At RCC GIS, we leverage big data, high-performance computing, and advanced algorithms to address complex geospatial challenges across various domains.

Our GeoAI Capabilities

Deep Learning for Spatial Analysis

We utilize advanced neural network architectures, including:

• Convolutional Neural Networks (CNNs) for image analysis
• Graph Neural Networks for network and connectivity analysis
• Recurrent Neural Networks (RNNs) for time-series geospatial data
• Transformer models for spatial-temporal predictions

Explainable GeoAI (XGeoAI)

Our research focuses on developing transparent and interpretable AI models for geospatial applications, ensuring that decision-makers can understand and trust the results of our analyses.

Generative AI for Geospatial Tasks

We explore the use of generative AI models, such as GeoGPT, for various applications:

• Map enrichment and feature extraction
• Synthetic data generation for training
• Spatiotemporal predictions and simulations

Integration with Big Data

Our GeoAI solutions are designed to handle large-scale spatial and temporal datasets, enabling high-resolution modeling and analysis for:

• Environmental epidemiology
• Exposure modeling
• Disease surveillance
• Urban dynamics

Featured Projects

Urban Resilience and Climate Adaptation

We use GeoAI to predict and manage urban resilience against extreme weather events, assess climate risks, and support disaster response planning. Our models help cities prepare for and adapt to climate change impacts.

Public Health and Epidemiology

Our GeoAI models analyze spatial and temporal data for:

• Disease mapping and hotspot detection
• Exposure modeling for environmental health
• Healthcare accessibility analysis
• Pandemic response planning

Environmental Monitoring

We leverage AI and machine learning techniques to analyze remote sensing data for:

• Deforestation monitoring and prediction
• Ocean eddy tracking
• Air quality assessment
• Land use/land cover change detection

Resources and Tools

Software and Libraries

• Deep Learning Frameworks: TensorFlow, PyTorch, PyTorch Lightning
• Geospatial AI Libraries: TorchGeo, Rasterio, GeoPandas, TensorFlow Extended (TFX)
• Cloud Platforms: Google Earth Engine, Microsoft Planetary Computer
• High-Performance Computing: Integration with UChicago’s Midway cluster

Training and Workshops

We offer various training opportunities:

1. Introduction to GeoAI

   • Fundamentals of machine learning for spatial data
   • Hands-on with geospatial Python libraries
   • Practical applications and case studies

2. Advanced Deep Learning for Remote Sensing

   • Image classification and segmentation
   • Object detection in satellite imagery
   • Time-series analysis of Earth observation data

3. Spatial Data Science with Python

   • Geospatial data manipulation
   • Spatial statistics and machine learning
   • Big data processing with Dask and GeoPandas

Getting Started

Prerequisites

• Basic knowledge of Python programming
• Familiarity with GIS concepts and tools
• Understanding of basic statistics and machine learning concepts (for advanced topics)

Recommended Learning Path

1. Complete our Introduction to Python for GIS
2. Take the Spatial Data Science with Python workshop
3. Explore our GeoAI starter projects on GitHub

Contact Us

Interested in incorporating GeoAI into your research? Our team of experts is here to help:

• Consultation: Discuss your project requirements
• Collaboration: Partner with us on research projects
• Training: Custom workshops for your team

Email us at gis@uchicago.edu to get started.