Post provided by Mateusz Iskrzyński
Want to quickly get a clear view of the food web you are modeling or analyzing? Have you thought about including his visualization in your work? Or maybe you wanted to show students or the general audience how interconnected food networks are in real ecosystems? Or just ask yourself how matter flows through an ecosystem?
To help you with this, Mateusz Iskrzyński and Łukasz Pawluczuk have developed an open source Python package featured in the recently published. Methods in Ecology and Evolution paper “Food Web Visualization: Heat Map, Interactive Graph and Animated Flow Network”. This package offers several complementary complementary methods of food network visualization: a heat map, an interactive graph, and an animated flow network, using colors, sizes, positions, zoom, and motion to accurately identify flows, their sizes and connection patterns. In this post, Mateusz describes the set of complementary complementary visualization methods that enable an overview of food networks within an ecosystem.
What does a food network look like?
Ecologists and field modelers carefully reconstruct who eats who and how much in an ecosystem. The food web produced is a weighted network that typically has dozens of nodes connected by streams that can vary up to ten orders of magnitude. Current flow values are really important for many properties of the food web, e.g. the cycle of matter in ecosystems. We were really surprised when we observed no correlation between its weightless simplification and proper evaluation.
When working with feed network data, we often lacked a satisfactory visualization tool that would allow us to quickly see the entire flow of matter allowing a clear identification of the nodes they connect to. To date, there has not been an open source library to draw weighted food networks with partial trophic levels clearly. Existing approaches are distributed across specific libraries and programming languages, and therefore they are challenging to use. They also miss an excellent opportunity to clearly display food nets.
Food networks are special. They are very hierarchical, which becomes apparent when the nodes are lined up according to their trophic level. A certain type of food requires specialization. Digesting plants or waste presents completely different challenges from carnivores. At the top of this come the numerous rankings of the organisms, leading to a degree of ‘trophic strata’.
The simplest and also traditional image of a food web designs the trophic levels of the nodes in their vertical positions. Our method creates an interactive graph that automatically optimizes the horizontal positions of the nodes. We gave users the ability to adjust algorithm settings or move nodes manually. The width of the arrows designs the size of the flow. The color of the nodes also encodes the trophic level. By hovering over a node, the user can display its basic properties. Selecting a node highlights all the flows associated with it.
The strengths of an interactive graphic come from allowing the viewer to track the flow of matter over subsequent links. However, all living organisms die and become waste, so the non-inclusion of such streams for the purpose of this presentation has significantly improved clarity.
Heat map for a clear identification of the network flow
The natural complexity of food nets makes them rarely easy to draw, as many of them consist of 50 or more knots. In a graph it becomes difficult to see which nodes are connected by a particular stream. A food web heat flow map plots the flow sizes that connect a row to a column with cell colors. It shows general patterns at a glance while maintaining accurate flow identification. The user can also display dietary proportions or normalized flows in other quantities such as their source biomass etc. as well as change the roles of columns and rows.
Attractive particle flow animations, such as Windy, inspired us to present matter moving through a food web. Particle motion helps visually identify their source and target, giving them an edge over conventional static representations. Particle density encodes feed network flows and node size compiles biomass reserves.
Multiple perspectives create a deeper understanding
We believe that combining multiple visualization approaches can improve understanding of an ecosystem and provide more relevant representations to the general public. Aesthetically pleasing and clear images that present empirical data help to communicate the importance of species interconnectedness and ecosystem complexity. We hope our package will help achieve these goals with minimal effort.
To read the full paper, ‘Food Internet Visualization: Heat Map, Interactive Graph and Animated Flow Network’, click here.
To watch the full video on YouTube about Mateusz newspaper here.