High-Quality Ultra-Compact Grid Layout of Grouped Networks

Abstract

Fig. 1: Grid systems in typographic layout, UI design and an example of our proposed grid layout for a power-graph. With graphic designers playing an increasing role in the design of user interfaces for phone, tablet and desktop operating systems, this traditional grid-based design aesthetic is becoming more popular in these media. A case in point is Microsoft's "Modern" interface which seeks to unify app-design across devices. This resurgence of the grid-design aesthetic in new media leads us to re-examine some of the aesthetic assumptions that have been made in designing layout methods for network diagrams.Abstract-Prior research into network layout has focused on fast heuristic techniques for layout of large networks, or complex multi-stage pipelines for higher quality layout of small graphs. Improvements to these pipeline techniques, especially for orthogonal-style layout, are difficult and practical results have been slight in recent years. Yet, as discussed in this paper, there remain significant issues in the quality of the layouts produced by these techniques, even for quite small networks. This is especially true when layout with additional grouping constraints is required. The first contribution of this paper is to investigate an ultra-compact, grid-like network layout aesthetic that is motivated by the grid arrangements that are used almost universally by designers in typographical layout. Since the time when these heuristic and pipeline-based graph-layout methods were conceived, generic technologies (MIP, CP and SAT) for solving combinatorial and mixed-integer optimization problems have improved massively. The second contribution of this paper is to reassess whether these techniques can be used for high-quality layout of small graphs. While they are fast enough for graphs of up to 50 nodes we found these methods do not scale up. Our third contribution is a large-neighborhood search meta-heuristic approach that is scalable to larger networks.

AVOCADO: Visualization of Workflow–Derived Data Provenance for Reproducible Biomedical Research

“…As in a spreadsheet, the largest node (e.g., an expanded workflow) defines the width for the remaining cells in the same column, which results in long edges to nodes in adjacent columns. Other layout approaches might create a more compact layout (e.g., [ YDG*15 ]), but must also consider the characteristics of the data provenance graph (see Section 2) and the interactive requirements to ensure fluid transitions when changing hierarchy levels. Since edge crossings decrease the readability of the graph, they are to be minimized.…”

Visualizing proportions and dissimilarities by Space-filling maps: A Large Neighborhood Search approach

“…The LNS metaheuristic was first proposed in [50]. It has been successfully applied in recent years to problems of different nature, for instance Vehicle Routing Problems, [16,40,48], Scheduling, [35,46] and Visualization [59]. Roughly speaking, LNS performs a search for good solutions in a neighborhood of a starting point.…”

Visualizing Group Structures in Graphs: A Survey

“…Such contours share the characteristics of set diagrams such as Euler diagrams. Contour shapes are versatile, for instance, rectangles [DGC*05, DHRMM13, HRD10, YDG*15] (Figures b.2 and b.3), circle sections [ET07] or circles [Hol06, KG06, NIS15] (Figure b.1), convex hulls [BPF14, ST08, WWY*15], arbitrary two‐dimensional curves or splines [BBT06, BD05, BT06, BT09b, DEKB*14, DvKSW12, EHKP14, GHK10, HGK10, HKV14, LQB12, VPF*14] (Figure c), or three‐dimensional bubbles [BD07, SBG00]. The GMap approach [GHK10, HGK10, HKV14] creates a map of contours that are adjacent to each other using a Voronoi tessellation.…”

“…To untangle overlapping contours, Henry Riche and Dwyer [HRD10] introduced two techniques for rectangular contour overlays: a splitting approach (groups with intersections are split up, drawn as non‐overlapping rectangular shapes, and linked by lines) and a duplication approach (Figure b.2) (groups are represented by overlaid rectangles and nodes contained in several groups are duplicated and linked visually). For disjoint hierarchical group structures , the contours or surfaces are nested to encode the hierarchical structure visually [BD05, BD07, DGC*05, DHRMM13, Hol06, KG06, SBG00, YDG*15] (e.g. Figure b.3).…”