VisWaterNet: A Python package for visualization of water distribution networks

Abstract

Water distribution systems (WDSs) are complex networks and are typically represented as graphs with multiple types of nodes (e.g., junctions, reservoirs, and tanks) and links (e.g., pipes, pumps, and valves). These network elements have both static characteristics, such as pipe diameter and nodal elevation, and time-varying properties, such as pipe flow rate and nodal pressure. Researchers and practitioners working with WDSs need to model, simulate, analyze, and visualize these systems. The visualization of data over network topology can lend spatial context and a better understanding of simulation results and network properties. This paper presents VisWaterNet, an open-source Python package that enables easy and customizable visualization of WDS static and dynamic properties. The package allows plotting node and link data by varying element size and color in a continuous or discrete manner. VisWaterNet also allows plotting of categorical data, custom data generated by the user, and animations to demonstrate the change in network properties over time. Statement of need Existing open-source tools for analyzing WDSs, such as EPANET (Rossman et al., 2020) and the Water Network Tool for Resilience Python package (WNTR) (Klise et al., 2017), allow users to perform extended period simulations and analyze system hydraulic performance under a range of different conditions. However, while these tools provide basic network visualization options, they are limited in their flexibility and capabilities. EPANET wrappers in other programming languages also offer few plotting functions; for example, an EPANET-interfacing toolkit developed for MATLAB (Eliades et al., 2016) only generates network layouts, and a similar R application (Arandia & Eck, 2018) generates basic network representations. On the other hand, popular software such as Cytoscape (Shannon et al., 2003) and Graphviz (Ellson et al., 2004) allow the easy integration of attribute data into sophisticated network plots, but it is challenging to apply them to provide meaningful results in the context of WDSs. VisWaterNet was developed to provide an accessible and multifunctional approach to WDS-specific network visualization. The package is accompanied by clear documentation and code examples in order to facilitate ease of adoption, and can prove to be a helpful tool for researchers and practitioners in the domain of WDS analysis. Users are welcome to further develop, improve and extend these open-source scripts.