LASPATED: A Library for the Analysis of Spatio-Temporal Discrete Data (User Manual)

This is the User Manual of the LASPATED library. This library is available on GitHub (at https://github.com/vguigues/LASPATED)) and provides a set of tools to analyze spatiotemporal data. A video tutorial for this library is available on Youtube. It is made of a Python package for time and space discretizations and of two packages (one in Matlab and one in C++) implementing the calibration of the probabilistic models for stochastic spatio-temporal data proposed in the companion paper arXiv:2203.16371v2.

💡 Research Summary

The paper presents a comprehensive user manual for LASPATED, a specialized software library engineered for the sophisticated analysis of spatiotemporal discrete data. As spatiotemporal phenomena—ranging from epidemic spreads to ecological shifts—are inherently complex and continuous, analyzing them requires a structured approach to discretization and probabilistic modeling. LASPATED addresses this challenge by providing a multi-language toolkit that integrates Python, C++, and Matlab, ensuring both ease of use and high-performance computational capabilities.

The architecture of LASPATED is strategically divided into functional modules tailored to different stages of the analytical workflow. The Python component serves as the primary interface for data preprocessing and discretization. It allows researchers to transform continuous spatial and temporal dimensions into discrete grids, a fundamental step for handling spatiotemporal datasets. By leveraging Python’s extensive ecosystem, the library facilitates seamless integration with existing data science workflows, making the initial stages of data preparation highly accessible.

The core computational heavy-lifting is handled by the C++ and Matlab packages. These modules are specifically designed to implement the calibration of probabilistic models for stochastic spatiotemporal data, as theoretically established in the companion paper (arXiv:2203.16371v2). Calibration is a critical and computationally intensive process involving the estimation of model parameters to align probabilistic predictions with observed discrete data. The C++ implementation provides the necessary computational efficiency for large-scale, high-performance tasks, while the Matlab package offers a robust environment for researchers accustomed to numerical analysis and engineering-centric modeling.

Beyond its technical implementation, LASPATED represents a bridge between theoretical stochastic modeling and practical scientific application. The library’s availability on GitHub ensures transparency and reproducibility, which are cornerstones of modern scientific research. Furthermore, the inclusion of supplementary educational resources, such as YouTube tutorials, lowers the barrier to entry for scientists across various disciplines, including ecology, epidemiology, and urban science. In essence, LASPATED provides a unified, end-to-end framework that guides users from the raw discretization of space and time to the advanced calibration of complex stochastic models, making it an indispensable asset for the spatiotemporal data science community.