PyLandStats: An open-source Pythonic library to compute landscape metrics

Abstract

Quantifying the spatial pattern of landscapes has become a common task of many studies in landscape ecology. Most of the existing software to compute landscape metrics is not well suited to be used in interactive environments such as Jupyter notebooks nor to be included as part of automated computational workflows. This article presents PyLandStats, an open-source Python library to compute landscape metrics within the scientific Python stack. The PyLandStats package provides a set of methods to undertake recurrent approaches to quantify landscape patterns, such as the analysis of the spatiotemporal patterns of land use/land cover change or gradient analysis. The implementation is based on the prevailing Python libraries for geospatial data analysis in a way that they can be forthwith integrated into complex computational workflows. Notably, the provided methods offer a large variety of options so that users can employ PyLandStats in the way that best supports their needs. The source code is publicly available, and is organized in a modular object-oriented structure that enhances its maintainability and extensibility.