Integrating Complex Geospatial Data into GenAI with GeoAI Chat Components

In an increasingly data-driven world, the sheer volume and complexity of information can be overwhelming, especially when it comes to geospatial data. Traditional access and extraction methods are time consuming and requires specialized training to get in a single form to sift and analyze datasets. Decision makers demand immediate, intuitive access to interconnected data, far beyond what a static map can provide.

Geospatial data is notoriously complex, often existing in various formats and systems. If the data can be made available to an AI environment to seamlessly connecting disparate data sources – maps, documents, and raw data – then the data can be ‘asked’ in a natural language. This requires various solutions as the insights derived from this data are often buried within lengthy reports, specialized databases or services and are inaccessible to those who need them most in a timely manner.

This is where the power of conversational AI meets the richness of geospatial data. Our GeoAI Chat Components are designed to empower users, regardless of their technical expertise, to unlock the full potential of their data. We are bridging this gap with our innovative GeoAI Chat Components, a suite of tools that transforms how users interact with geospatial data, making it as simple as having a conversation.

Our goal was to create components that could not only understand natural language queries but also intelligently retrieve, process, and present geospatial information from various sources in an understandable and actionable format.The four components for Geospatial Extraction and Integration from:

  • GeoAI Map Explorer: Natural Language and Fuzzy Logic calls to geospatial services (Fuzzy Logic/NLP)
  • GeoAI Documents Explorer: Documents in Unstructured Data Lake (Data Lake)
  • GeoAI Data Explorer: Data via Retrieval-Augmented Generation (RAG)
  • LLM Add-on & Orchestrator: Traditional Integration with Large Language Model (LLM) and AI Agents

The GeoAI Chat Components leverage LLMs to act as a sophisticated interpreter, translating human language into actionable commands for geospatial analysis and data retrieval. This allows users to simply ask questions like: "Show me all active wildfires in California and their proximity to residential areas," or "Summarize the environmental impact assessment for the proposed development in Block 7." The LLM then orchestrates the necessary actions across our specialized components to deliver a comprehensive answer.

A Suite of Intelligent Geospatial Explorers

#1 GeoAI Maps Explorer

The GeoAI Map Explorer transforms how users interact with maps. Instead of clicking through layers and tools, users can simply ask questions about geographic features, and the system dynamically adjusts the map display to provide the requested information.

  • Real-world Use Case: An emergency management team needs to understand the current evacuation zones and the location of available shelters during a hurricane. They can ask, "Show me evacuation zones for Hurricane Zeta and highlight available shelters." The GeoAI Map Explorer would then display this information directly on the map, allowing for immediate visual analysis and informed decision-making.
  • Capability Benefits: 
    • Intuitive Map Interaction: Navigate and query maps using natural language, eliminating the need for complex GIS software training.
    • Dynamic Data Overlay: Overlay various geospatial datasets on demand, such as weather patterns, population density, or infrastructure.
    • Enhanced Situational Awareness: Quickly visualize critical information to support rapid response and planning.

Personalized Map Views: Users can customize map displays based on their specific needs and interests.

Test Harness UI to add any feature service and ask questions using natural language.

JSON Inspector shows how call is translated to AI

 Query results show extracted highlighted data (in this case roads that fit the criteria) as well as table




#2 GeoAI Documents Explorer

The GeoAI Documents Explorer is designed to extract, summarize, and answer questions from large volumes of unstructured textual data, such as reports, permits, and historical documents, connecting them to their geospatial context.

  • Real-world Use Case: A city planner needs to review environmental impact reports for a new construction project. Instead of reading hundreds of pages, they can ask, "Summarize the key environmental concerns identified in the permit applications for the new downtown high-rise project." The Documents Explorer would then provide a concise summary, highlighting relevant sections and potential geospatial implications.
  • Capability Benefits:
    • Intelligent Document Search & Summarization: Quickly find specific information within vast document repositories and generate concise summaries.
    • Contextual Understanding: Link document content to geospatial locations, providing a richer understanding of data.
    • Time Savings: Drastically reduce the time spent manually reviewing documents.

Improved Decision-Making: Access critical insights from documents that might otherwise remain undiscovered.

Data Lake Explorer can select specific documents and ask questions in natural language

#3 GeoAI Data Explorer

Conversational Access to Structured Data;

The GeoAI Data Explorer empowers users to query and analyze structured geospatial datasets through natural language, eliminating the need for complex database queries or statistical software.

  • Real-world Use Case: A researcher is studying urban growth patterns and wants to know the population change in specific zip codes over the last decade, along with the corresponding changes in commercial property values. They can ask, "What was the percentage population change in zip codes 84094 and 84095 between 2010 and 2020, and what was the average commercial property value increase in those areas during the same period?" The GeoAI Data Explorer would then retrieve and present this statistical information.
  • Capability Benefits: 
    • Ad-hoc Data Analysis: Perform complex queries on structured data without writing code or using specialized tools.
    • Data Aggregation and Reporting: Generate custom reports and visualizations based on specific data points.
    • Democratization of Data: Make powerful data analysis accessible to a wider range of users.

Rapid Insight Generation: Quickly identify trends, anomalies, and correlations within datasets.

Easy upload, configuration file, or even STAC Explorer can guide adding data sources to RAG.

 

After loading the RAG with the layers you want to include, a user can then ask a natural language question that cuts across those data sources

 

#4 GeoAI LLM Orchestrator

4. The LLM Add-on is an additional embedded AI connection that can be utilized for non-geospatial related queries and provides additional contextual information to the geospatial operations without needing to leave the interface.

  • This is where the orchestration comes to fruition
  • Call traditional LLM asks
  • Use AI Agents to call and outputs the other 3 components

Bringing it all together with a standard ChatBot interface. The demo below shows the components separately to show how it is managed, yet can easily be brought into the GeoAI Chatbot LLM as AI Agent calls based on desired output - show on a map, show as data, explore document outputs, etc.

Putting it all together: The GeoAI Chat were built as a series of components which can be integrated into the LLM ChatBot. Watch this video to see a full review of all components:

What's Next in the Workbench

The GeoWorkbench has already proven its value as a platform for collaboration, training, and project development. Our vision for the future is to expand its capabilities even further by:

  • Continue to use it to help jumpstart Labs, POCs, and development work for clients where approved.
  • Creating customized GWB images for specific GIS and data science tasks.
  • Integrating the GWB directly into client business processes.
  • Using it as a sandbox to explore and develop new server technologies and custom extensions.

The GeoWorkbench is more than just a tool; it's a core component of our strategy to foster innovation and empower our team with the resources they need to succeed.