Exploring Graphs at Scale (EGAS) 2015
Big Graph Visual Analytics Challenges and Opportunities
IEEE VIS 2015 Workshop
Monday, October 26, 2015 - Chicago, IL

About


IEEE EGAS 2015 will explore the technical challenges and technology development opportunities of graph visual analytics found in the big data era with the goal of solving big problems with big graph data. Today’s graph problems are increasingly multifaceted and multidisciplinary in nature. Many cutting-edge R&D efforts are conducted independently in disparate domains such as bioinformatics, cybersecurity, and predictive machine learning. Although technology transfers in big graph visualization and visual analytics are recognized and growing, little progress has been made in establishing a community strategy for sharing and building knowledge.

We invite researchers and practitioners with different interests to participate at IEEE EGAS 2015 by submitting technical or position papers and, if accepted, presenting their ideas at the workshop co-located at IEEE VIS 2015. A technical paper aims to discuss preliminary but promising results, whereas a position paper is meant to share visionary ideas and/or unsolved challenges that the paper authors have encountered.

We agree that the data size that seems big today is different from what seemed big only a few years ago. While the workshop doesn’t specify upper or lower bounds on the graph’s size, we are particularly interested in emerging problems that challenge conventional wisdom in computation and interaction brought by the latest social-scale or web-scale graphs.

Some of these pressing challenges are:

  • What if graph analytics did not start with visualization? What if that happened downstream in the workflow?
  • Some graphs have attributes on the edges; some have multiple edges between two vertices. In what way does the “flavor” of graph affect the analytics that are possible?
  • Some edge attributes naturally form class hierarchies. How can this help the visual analytics process?
  • For those graphs with temporal and/or geo-spatial information associated with the vertices and/or edges, how can this information improve the analytics?
  • What metrics or techniques are best at showing a summary of the graph data?
  • If I have a system with only 1TB of memory, how can I perform visual analytics on a graph that requires 20TB to represent it?
  • In database parlance, Extract, Transform, and Load (ETL) may account for over 80% of the effort required to manipulate data. Are there ways to improve ETL using graph visualization and analytics techniques? Can better ETL be used to improve big graph visual analytics?
  • Can 10,000 users concurrently run visual analytics software aimed at the same graph data source?
  • Can big graph data be viewed across a network? For example, can a user in Europe run a visual analytics tool on data stored in South America?

Selected References

P Burkhardt and C Waring, An NSA Big Graph Experiment, Technical Report NSA-RD-2013-056002v1, U.S. National Security Agency, May 20, 2013.

P Burkhardt, “Big Graphs,” The Next Wave, 20(4), 2014.

VG Castellana, A Morari, J Weaver, A Tumeo, D Haglin, O Ville, and J Feo, “In-Memory Graph Databases for Web-Scale Data,” IEEE Computer, 48(3):24-35, Mar 2015.

D Ediger, K Jiang, EJ Riedy, and DA Bader, “GraphCT: Multithreaded Algorithms for Massive Graph Analysis,” IEEE Transactions on Parallel & Distributed Systems, 2012.

R Rohrer, CL Paul, and B Nebesh, “Visual Analytics for Big Data,” The Next Wave, 20(4), 2014.

A Morari, VG Castellana, O Villa, A Tumeo, J Weaver, D Haglin, J Feo, and S Choudhury, “Scaling Semantic Graph Databases in Size and Performance,” IEEE Micro, 34(4):16-26, Jul 2014.

PC Wong, H-W Shen, C Chen, C Johnson, and R Ross, “Top Ten Challenges in Extreme-Scale Visual Analytics,” IEEE Computer Graphics and Applications, 32(4):63-67, IEEE CS Press, Jul 2012.

Details

Important Dates

Submission Deadline
Friday, August 14, 2015

Author Notification
Friday, Sept 18, 2015

Workshop Date
8:00am - Noon
Monday, Oct 26, 2015

Submission

IEEE EGAS 2015 solicits 3-to-4 page technical or position papers. Authors must follow the IEEE TVCG formatting guidelines and submit pdf files via EASYCHAIR. Submissions will be reviewed by 2-3 domain experts. Accepted papers will be presented at the workshop and posted online (further information TBA). Authors will maintain the copyright to their work, which allows them to submit to or present at a different publication venues in the future.

Keynote Speaker
"Graph it! Vis it!"

Paul Burkhardt
U.S. National Security Agency

Paul Burkhardt is a computer science researcher in the Research Directorate at NSA. He received his PhD from the University of Illinois at Urbana-Champaign. His current research interests are primarily focused on graph algorithms and Big Data analytics.

Workshop Program

8:30 - 8:35 Pak Chung Wong Opening, introduction
8:35 - 9:15 Paul Burkhardt Keynote: "Graph it! Vis it!" (Presentation)
9:15 - 9:30 Scott Langevin, David Jonker, David Giesbrecht and Michael Crouch Multi-Scale Community Visualization of Massive Graph Data (Presentation)
9:30 - 9:45 Lorne Leonard, Kamesh Madduri and Chris Duffy Graph-based Analysis for Large-scale Hydrological Modelling (Presentation)
9:45 - 10:00 Michael Burch Visualizing Large Dynamic Digraphs (Presentation)
10:00 - 10:05 Dongyu Liu, Fangzhou Guo, Bowen Deng, Yingcai Wu and Huamin Qu egoComp: A Node-link Based Technique for Visual Comparison of Ego-network (Presentation)
10:05 - 10:10 Guohao Zhang, Peter Kochunov, Elliot Hong, Keqin Wu, Hamish Carr and Jian Chen A Semantic Contour Tree Approach for Visual Comparison of Brain White Matter Connectivities in Cohorts
10:10 - 10:30 Coffee Break Coffee Break
10:30 - 10:45 William Longabaugh Using Linear Visualization to Explore Large Graphs (Presentation)
10:45 - 11:00 Joachim Giesen, Philipp Lucas, Claudia Dahl, Klaus Mueller and Shenghui Cheng Exploring the Distribution of Local Neighborhood Structures in Large Networks (Presentation)
11:00 - 11:15 Alexander Garbarino, Zachary Garbarino, Liang Sun, Carl Schmidt and Jian Chen VisGumbo, VisMirror, VisCut: Interactive Narrative Strategies for Large Biological Pathway Comparisons
11:15 - 11:30 Keqin Wu, Liang Sun, Carl Schmidt and Jian Chen A Graph Query Algebra for Biological Pathway Exploration
11:30 - 11:45 Fangyan Zhang, Song Zhang, Pak Chung Wong, J. Edward Swan and T.J. Jankun-Kelly A Visual and Statistical Benchmark for Graph Sampling Methods (Presentation)
11:45 - 12:10 David Haglin Townhall meeting, closing

Workshop Organizers

Pak Chung Wong

Pacific Northwest National Laboratory

Pak Chung Wong is a chief scientist and project manager at PNNL. He has spent the last decade delivering working solutions to big-graph problems in areas such as social networks, telecommunication networks, cyber networks, and power grids. More recently, his research has focused on extreme-scale visual analytics as applied to exa-scale scientific datasets and web-scale graphs with up to a trillion edges. Wong serves as an Associate Editor-in-Chief of IEEE Computer Graphics and Applications and an Associate Editor of Information Visualization. In the past, he chaired IEEE VisWeek, IEEE InfoVis, IEEE VAST, and SPIE VDA conferences.

David Haglin

Pacific Northwest National Laboratory

David Haglin is the chief scientist for High Performance Data Analytics at PNNL where he leads a team of researchers working on large, sparse, irregular computational problems that are typically best characterized as graph computations. One of the key contributions of this program is a software ecosystem called Graph Engine for Multithreaded Systems (GEMS). This software supports the flexible attributed multi-graph data model. His earlier work focused on graph algorithms, machine learning, and high-performance computing. He has served on the program committee for IEEE IPDPS and is recognized as a Senior Member of IEEE and a Senior Member of ACM.

David A. Bader

Georgia Institute of Technology

David A. Bader is a Full Professor and Chair of the School of Computational Science and Engineering, College of Computing, at Georgia Institute of Technology, and Executive Director of High Performance Computing. His interests are at the intersection of high-performance computing and real-world applications, including computational biology and genomics and massive-scale data analytics. He is a Fellow of the IEEE and AAAS, a NSF CAREER Award recipient, and a co-founder of the Graph500 List for benchmarking “Big Data” computing platforms. Bader is recognized as a "RockStar" of High Performance Computing by InsideHPC and as HPCwire's People to Watch in 2012 and 2014.

David Trimm

U.S. Department of Defense

David Trimm received his PhD from the University of Maryland in 2012. As a Department of Defense employee, Dr. Trimm has over twenty years of experience working the cybersecurity mission. In this role, Dr. Trimm researches advanced tradecraft to include visual analytics, graph algorithms, and machine learning techniques.

Workshop Committee

  • Seung-Hee Bae, University of Washington
  • Jessica Chang, DoD
  • Polo Chau, Georgia Tech
  • Abon Chaudhuri, WalmartLabs
  • John Feo, Context Relevant
  • John Johnson, Pacific Northwest National Laboratory
  • Randy Rohrer, DoD
  • Song Zhang, Mississippi State University