Skip to content

What is Geospatial Studio?

The IBM Geospatial Exploration and Orchestration Studio (Geospatial Studio) is an integrated platform for fine-tuning, inference, and orchestration of geospatial AI models. It makes working with geospatial data and AI accessible to everyone, from researchers to developers.

🎯 Core Purpose

Geospatial Studio addresses a critical challenge: making geospatial AI accessible and scalable. While satellite imagery and Earth observation data are abundant, extracting insights requires specialized knowledge and infrastructure. Geospatial Studio bridges this gap.

🏗️ Platform Components

The platform combines three interaction modes:

1. No-Code UI

A web-based interface for visual interaction with the platform. Perfect for: - Exploring datasets and models - Running inference visually - Monitoring training progress - Visualizing results on interactive maps

2. Low-Code SDK

A Python SDK for programmatic access. Ideal for: - Jupyter notebook workflows - Automated pipelines - Custom integrations - Batch processing

3. RESTful APIs

Direct API access for: - System integration - Custom applications - CI/CD pipelines - Microservices architecture

🔧 What Can You Do?

Dataset Management

  • Onboard training datasets from various sources
  • Validate data quality and format
  • Prepare data for model training
  • Organize datasets in a catalog

Model Fine-Tuning

  • Customize foundation models for specific tasks
  • Configure training parameters
  • Monitor training progress with MLflow
  • Evaluate model performance

Inference at Scale

  • Run models on large geospatial datasets
  • Process satellite imagery automatically
  • Visualize results on interactive maps
  • Export outputs for further analysis

Orchestration

  • Automate end-to-end workflows
  • Chain multiple processing steps
  • Schedule recurring tasks
  • Integrate with existing systems

🌟 Key Features

Built on Proven Technology

Geospatial Studio leverages the TerraStackAI ecosystem:

  • TerraTorch - Model fine-tuning and inference framework
  • TerraKit - Geospatial data search, query, and processing
  • Iterate - Hyperparameter optimization

Enterprise-Ready Deployment

  • On-premises or cloud - Deploy on Red Hat OpenShift or Kubernetes
  • Flexible configuration - Choose in-cluster or external cloud services
  • Scalable - Handle large-scale processing workloads
  • Secure - OAuth2 authentication and RBAC
  • Production-grade - High availability and monitoring

Deployment Flexibility

Local Deployment: - Fixed in-cluster services (PostgreSQL, MinIO, Keycloak) - Ideal for learning, testing, and development - Quick setup with Lima VM

Cluster Deployment: - Choose between in-cluster or external cloud-managed services - Database: In-cluster PostgreSQL OR IBM Cloud, AWS RDS, Azure, GCP - Storage: In-cluster MinIO OR IBM COS, AWS S3, Azure Blob, GCP Storage - Auth: In-cluster Keycloak OR IBM Verify, Okta, Azure AD - Perfect for production and enterprise deployments

Comprehensive Tooling

  • MLflow - Experiment tracking and model registry
  • GeoServer - Geospatial data visualization
  • PostgreSQL - Metadata storage
  • MinIO - S3-compatible object storage
  • Redis - Caching and message queuing

🎓 Who Is It For?

Data Scientists

  • Fine-tune models for specific geospatial tasks
  • Experiment with different architectures
  • Track and compare model performance
  • Deploy models for inference

Researchers

  • Process large satellite imagery datasets
  • Analyze Earth observation data
  • Publish reproducible results
  • Collaborate on geospatial AI projects

Developers

  • Build geospatial applications
  • Integrate AI capabilities via APIs
  • Create custom workflows
  • Automate data processing

Domain Experts

  • Use fine-tuned models without coding
  • Visualize results on maps
  • Extract insights from satellite data
  • Monitor environmental changes

🌍 Real-World Applications

Environmental Monitoring

  • Track deforestation and reforestation
  • Monitor water resources
  • Assess ecosystem health
  • Measure carbon sequestration

Disaster Response

  • Map flood extent
  • Detect wildfire burn scars
  • Assess infrastructure damage
  • Support emergency planning

Climate Analysis

  • Downscale climate models
  • Analyze land use changes
  • Monitor glaciers and ice sheets
  • Study urban heat islands

Agriculture

  • Crop health monitoring
  • Yield prediction
  • Irrigation optimization
  • Pest detection

Urban Planning

  • Building detection
  • Infrastructure mapping
  • Population estimation
  • Land use classification

🔄 Complete ML Lifecycle

Geospatial Studio supports the entire machine learning lifecycle:

graph LR
    A[Data Collection] --> B[Data Preparation]
    B --> C[Model Training]
    C --> D[Model Evaluation]
    D --> E[Model Deployment]
    E --> F[Inference]
    F --> G[Visualization]
    G --> H[Insights]
    H --> A

    style A fill:#0f62fe,stroke:#fff,stroke-width:2px,color:#fff
    style B fill:#8a3ffc,stroke:#fff,stroke-width:2px,color:#fff
    style C fill:#33b1ff,stroke:#fff,stroke-width:2px,color:#fff
    style D fill:#007d79,stroke:#fff,stroke-width:2px,color:#fff
    style E fill:#ff7eb6,stroke:#fff,stroke-width:2px,color:#fff
    style F fill:#fa4d56,stroke:#fff,stroke-width:2px,color:#fff
    style G fill:#42be65,stroke:#fff,stroke-width:2px,color:#fff
    style H fill:#f1c21b,stroke:#000,stroke-width:2px,color:#000

💡 Why Geospatial Studio?

Accessibility

  • No specialized knowledge required - Use fine-tuned models out of the box
  • Guided workflows - Step-by-step processes for common tasks
  • Visual interface - No coding required for basic operations

Flexibility

  • Multiple interfaces - UI, SDK, or API based on your needs
  • Customizable - Fine-tune models for your specific use case
  • Extensible - Integrate with existing tools and workflows

Scalability

  • Cloud-native - Kubernetes-based architecture
  • GPU acceleration - Fast training and inference
  • Distributed processing - Handle large datasets efficiently

Reproducibility

  • Experiment tracking - MLflow integration
  • Version control - Track models and datasets
  • Documented workflows - Share and reproduce results

🚀 Getting Started

Ready to start using Geospatial Studio? Here's what you'll learn in this workshop:

  1. Deploy the platform in your environment
  2. Navigate the UI and understand components
  3. Run inference with fine-tuned models
  4. Onboard datasets for training
  5. Fine-tune models for specific tasks
  6. Execute end-to-end workflows

📚 Learn More

← Back to Welcome Next: Architecture →