Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

MobilityDuck: Mobility Data Management with DuckDB

Nhu Ngoc Hoang, Ngoc Hoa Pham, Viet Phuong Hoang, Esteban Zimányi

TL;DR

MobilityDuck presents the first DuckDB extension that embeds the MEOS spatiotemporal engine to provide native spatiotemporal types and trajectory operators in an in-memory analytical environment. By connecting a lightweight, in-process analytics engine with MEOS, MobilityDuck preserves expressive spatiotemporal queries while leveraging DuckDB’s vectorized execution and Python interoperability. The BerlinMOD-Hanoi benchmark demonstrates that MobilityDuck delivers competitive, often superior performance compared with MobilityDB across multiple queries and scales, validating its practicality for urban mobility analytics. The work lays groundwork for deeper MEOS/Spatial integration, automated type bindings, and expanded indexing, enabling a unified, high-performance platform for spatiotemporal analytics in embedded analytics pipelines.

Abstract

The analytics of spatiotemporal data is increasingly important for mobility analytics. Despite extensive research on moving object databases (MODs), few systems are ready on production or lightweight enough for analytics. MobilityDB is a notable system that extends PostgreSQL with spatiotemporal data, but it inherits complexity of the architecture as well. In this paper, we present MobilityDuck, a DuckDB extension that integrates the MEOS library to provide support spatiotemporal and other temporal data types in DuckDB. MobilityDuck leverages DuckDB's lightweight, columnar, in-memory executable properties to deliver efficient analytics. To the best of our knowledge, no existing in-memory or embedded analytical system offers native spatiotemporal types and continuous trajectory operators as MobilityDuck does. We evaluate MobilityDuck using the BerlinMOD-Hanoi benchmark dataset and compare its performance to MobilityDB. Our results show that MobilityDuck preserves the expressiveness of spatiotemporal queries while benefiting from DuckDB's in-memory, columnar architecture.

MobilityDuck: Mobility Data Management with DuckDB

TL;DR

MobilityDuck presents the first DuckDB extension that embeds the MEOS spatiotemporal engine to provide native spatiotemporal types and trajectory operators in an in-memory analytical environment. By connecting a lightweight, in-process analytics engine with MEOS, MobilityDuck preserves expressive spatiotemporal queries while leveraging DuckDB’s vectorized execution and Python interoperability. The BerlinMOD-Hanoi benchmark demonstrates that MobilityDuck delivers competitive, often superior performance compared with MobilityDB across multiple queries and scales, validating its practicality for urban mobility analytics. The work lays groundwork for deeper MEOS/Spatial integration, automated type bindings, and expanded indexing, enabling a unified, high-performance platform for spatiotemporal analytics in embedded analytics pipelines.

Abstract

The analytics of spatiotemporal data is increasingly important for mobility analytics. Despite extensive research on moving object databases (MODs), few systems are ready on production or lightweight enough for analytics. MobilityDB is a notable system that extends PostgreSQL with spatiotemporal data, but it inherits complexity of the architecture as well. In this paper, we present MobilityDuck, a DuckDB extension that integrates the MEOS library to provide support spatiotemporal and other temporal data types in DuckDB. MobilityDuck leverages DuckDB's lightweight, columnar, in-memory executable properties to deliver efficient analytics. To the best of our knowledge, no existing in-memory or embedded analytical system offers native spatiotemporal types and continuous trajectory operators as MobilityDuck does. We evaluate MobilityDuck using the BerlinMOD-Hanoi benchmark dataset and compare its performance to MobilityDB. Our results show that MobilityDuck preserves the expressiveness of spatiotemporal queries while benefiting from DuckDB's in-memory, columnar architecture.

Paper Structure

This paper contains 31 sections, 12 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Background and Related Work
  3. Spatiotemporal Data Management
  4. The MEOS Library
  5. Benchmarks for Moving Object Databases
  6. DuckDB and In-process Analytics
  7. MobilityDuck: Architecture and Implementation
  8. Design Goals
  9. System Architecture
  10. Type System and Registration
  11. Registration of Functions and Operators
  12. Sample Queries
  13. Indexing System
  14. Index Registration
  15. Index Construction
  16. ...and 16 more sections

Figures (12)

  • Figure 1: Execution plan of query
  • Figure 2: R-tree index scan versus sequential scan of MobilityDuck and DuckDB on different scale factors
  • Figure 3: Visualization of animated synthetic trips in Hanoi generated by BerlinMOD-Hanoi.
  • Figure 4: Hanoi administrative regions used in BerlinMOD-Hanoi.
  • Figure 5: Visualization of a synthetic trip in Hanoi generated by BerlinMOD-Hanoi (QGIS+MOVE plugin).
  • ...and 7 more figures