An Innovative Food Web Display

CPSC 533C Project Proposal

November 4, 2005

Zhangbo(Zephyr) Liu zephyr@cs.ubc.ca

Brief Description

This project ties in with one of the projects I'm doing for CPSC 544-Human-Computer Interaction course. Our project will involve prototyping, evaluating, redesigning and implementing a food web visualization diagram within a complex ecological modeling software suite called Ecopath with Ecosim (EwE). We'll redesign the system together and then I'll do the implementation work while the other two members, Sherman Lai and Paul Ralph, concentrate on evaluation.

Setting and Problem

One of the functions of Ecopath involves showing trophic relationships (essentially what eats what and in what amounts, in an ecosystem) using food web diagrams. Figure 1 shows the kind of food web diagram that might be shown in an elementary school science class. Although this diagram is easy to read, it lacks detail. First, it shows groups of organisms at inconsistent and generally coarse levels of granularity. Second, it shows only what each organism eats and not how much. Third, it does not show cannibalistic relationships.

Figure 1: A Simple Food Web ¨C a small and basic food web diagram which does not show granularity, amount of interaction, and possible cannibalism

Compare this with the Ecopath diagram shown in Figure 2. For a target ecosystem at a particular time, the Ecopath food web diagram shows:

Figure 2: Ecopath Food Web Diagram ¨C demonstrates a more complicated food web diagram showing all the necessary information

Users have complained that beyond approximately 20 groups the display becomes incomprehensible. Figure 2 shows a 17-group food web. Due to the number of organisms and even groups of organisms in real-life ecosystems, models can have upwards of 200 groups. It seems clear that the capacity to display a more complex food web intelligibly would be a beneficial improvement to Ecopath.

Analysis

Ecopath's target users are scientists who publish papers describing ecosystems, more specifically trophic interactions. These scientists use Ecopath's food web diagrams to take a snapshot of the ecosystem for inclusion in their papers. Our goal then, is to increase the number of organism groups that can be clearly displayed on a printable diagram generated by Ecopath.

The necessary condition of representation on a two-dimensional medium, i.e. paper, constrains the useful depiction of a food web diagram in a variety of ways:

Three basic approaches can be applied to a food web diagram made unreadable by the amount of information displayed:

Given that users complain about not being able to display enough groups, we infer that displaying fewer groups is not a practical option. That is, unless we display groups at varying levels of abstraction. The second option is dubious, but we will inquire about it during the task analysis. The third option is likely the most desirable approach.

Most of the time, food webs have more organisms, both in total number and number of species, at the bottom than at the top. For example, the number of different birds that eat insects is much lower than the number of different insects eaten by birds. This causes food web diagrams to become more congested at the bottom than at the top. To fit more into the diagram, better use must be made of the remaining room at the top. However, the vertical distribution of groups in the diagram indicates trophic levels; thus, indiscriminately moving low-trophic-level groups to the top of the diagram would likely cause confusion and reduce utility.

Proposed Solutions

We propose to improve the readability of the diagram by restructuring the tree-type display into a circular display [Bertin, 83] . The center of the circle would represent the highest trophic level with concentric circles indicating progressively lower levels. This should allow displays of larger food webs in the same area.

We have in mind several further modifications that might increase the utility of the food web diagram (listed below). We expect more ideas to arise from discussions with users. Clearly, implementing all of these ideas is far beyond what is doable in the time constraints of this project; therefore, we will only implement the modifications that appear the most promising from the users' perspective. This could be one complex modification, or several simpler modifications.

Figure 3: Circular graph

Groups are arranged on concentric rings. Each ring represents a specific trophic level. When users click on one group (node), a pop-up window may show all existed information of this group such as name, biomass and trophic level. At the same time, all interactions between the selected group and other groups will be highlighted and the strength of trophic relationship is indicated by the thickness of the lines. See Figure 4

Figure 4: More information are shown when specific group is selected

Given that some users still prefer tree graph, we provide a button on the top right of the screen. Users can freely transfer to tree graph by clicking this button and make a compare. See Figure 5.

Figure 5: Tree-like graph

Proposed Implementation Approach

We will implement the interface using Microsoft Visual Basic.Net 2005 beta. It will run on Windows systems. We will start from scratch on the implementation and it will be integrated into current version of EwE.

Milestones

Before Nov. 4th	Early prototyping, user interview and analysis
Nov. 4th	Proposal submitted
Nov. 16th	Project updates, beta 1 released
Dec. 10th	Beta 2 released, start on writing report
Dec. 18th	Report finished
Dec. 19th	Final presentation

Referrences:

[Bertin, 83] Bertin, J. (1983) Semiology of Graphics, Diagrams, Networks, Maps, University of Wisconsin Press, Madison, WI.

[Furnas , 1986] Generalized fisheye views. In Proceedings of CHI'86, 16-23.

[Lamping et al., 1995] Lamping, J., Rao, R., Pirolli; P. (1995) A focus+context technique based on hyperbolic geometry for visualizing large hierarchies Conference proceedings on Humanfactors in computing systems, 1995, 401-408.