Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Flexible Architecture for Web-based GIS Applications using Docker and Graph Databases

Yves Annanias, Daniel Wiegreffe

TL;DR

The paper tackles the challenge of managing highly heterogeneous geospatial data for regional planning by proposing a flexible, Docker-based microservices architecture that separates frontend, backend, and a graph database. It leverages a graph model (Neo4j) to link entities such as projects, people, companies, and documents, enabling efficient cross-data queries via Cypher and supporting incremental data modeling. The system integrates advanced visualization (coordinated maps, heatmap-like aggregation) with document processing (active learning for statement extraction and PDF linking) and rich data sources (weather, sensors, LIDAR via Potree). This architecture enables scalable, extensible web-based GIS workflows for renaturation projects, reducing manual data linkage and facilitating insights across diverse data types. Practical lessons emphasize modularity, data exchange standards, and the value of a unified client entry point for enterprise deployment.

Abstract

Regional planning processes and associated redevelopment projects can be complex due to the vast amount of diverse data involved. However, all of this data shares a common geographical reference, especially in the renaturation of former open-cast mining areas. To ensure safety, it is crucial to maintain a comprehensive overview of the interrelated data and draw accurate conclusions. This requires special tools and can be a very time-consuming process. A geographical information system (GIS) is well-suited for this purpose, but even a GIS has limitations when dealing with multiple data types and sources. Additional tools are often necessary to process and view all the data, which can complicate the planning process. Our paper describes a system architecture that addresses the aforementioned issues and provides a simple, yet flexible tool for these activities. The architecture is based on microservices using Docker and is divided into a backend and a frontend. The backend simplifies and generalizes the integration of different data types, while a graph database is used to link relevant data and reveal potential new relationships between them. Finally, a modern web frontend displays the data and relationships.

A Flexible Architecture for Web-based GIS Applications using Docker and Graph Databases

TL;DR

The paper tackles the challenge of managing highly heterogeneous geospatial data for regional planning by proposing a flexible, Docker-based microservices architecture that separates frontend, backend, and a graph database. It leverages a graph model (Neo4j) to link entities such as projects, people, companies, and documents, enabling efficient cross-data queries via Cypher and supporting incremental data modeling. The system integrates advanced visualization (coordinated maps, heatmap-like aggregation) with document processing (active learning for statement extraction and PDF linking) and rich data sources (weather, sensors, LIDAR via Potree). This architecture enables scalable, extensible web-based GIS workflows for renaturation projects, reducing manual data linkage and facilitating insights across diverse data types. Practical lessons emphasize modularity, data exchange standards, and the value of a unified client entry point for enterprise deployment.

Abstract

Regional planning processes and associated redevelopment projects can be complex due to the vast amount of diverse data involved. However, all of this data shares a common geographical reference, especially in the renaturation of former open-cast mining areas. To ensure safety, it is crucial to maintain a comprehensive overview of the interrelated data and draw accurate conclusions. This requires special tools and can be a very time-consuming process. A geographical information system (GIS) is well-suited for this purpose, but even a GIS has limitations when dealing with multiple data types and sources. Additional tools are often necessary to process and view all the data, which can complicate the planning process. Our paper describes a system architecture that addresses the aforementioned issues and provides a simple, yet flexible tool for these activities. The architecture is based on microservices using Docker and is divided into a backend and a frontend. The backend simplifies and generalizes the integration of different data types, while a graph database is used to link relevant data and reveal potential new relationships between them. Finally, a modern web frontend displays the data and relationships.
Paper Structure (9 sections, 4 figures)

This paper contains 9 sections, 4 figures.

Table of Contents

  1. Introduction
  2. Related Work
  3. Functionalities of the System
  4. System Architecture
  5. Frontend
  6. Backend
  7. Data Linkage
  8. Lessons Learned
  9. Future Work & Conclusion

Figures (4)

  • Figure 1: Interlinked maps and information panels of the frontend. A) Projects with valid weather restrictions. B) Different feature categories are aggregated into a heatmap visualization. C) LIDAR data for the region. D) Projects with time restrictions grouped by month. E) Measured environmental data. F) PDF Viewer with marked sentence. G) Text data indicating a current weather restriction.
  • Figure 2: Overview of the document processing using different containers for specific sub-tasks.
  • Figure 3: Example graph showing how entities (persons / companies) extracted from the documents, are related to different projects.
  • Figure 4: Weather restrictions for an area can be checked using corresponding sensors nearby, even if the specified area does not have such a restriction directly (red path).