Note: This scenario and all the people, places, groups, technologies, contained therein are fictitious. Any resemblance to real people, places, groups, or technologies is purely coincidental.


Analysts at FishEye International are committed to stopping illegal, unreported, and unregulated (IUU) fishing and protecting marine species that are affected by it. As part of their work, FishEye collects online news articles about fishing, marine industry, and international maritime trade. To facilitate their analysis, FishEye uses a natural language processing tool to extract the names of entities (people and businesses) and the relationships between them. The extracted information is transformed into a knowledge graph that analysts would like to use to investigate illegal fishing. One of FishEye’s tasks is investigating tips they receive about possible IUU involvement. To push their investigations forward, FishEye’s analysts need to quickly see, understand, and explore the context around each tip. If they can connect an entity identified in a tip to illegal fishing, they may be able to escalate their investigation. Helpfully, FishEye has included some of their own illegal fishing case studies in the database so the organization and the companies they have investigated appear in the graph.

However, simply showing all information related to a suspected entity will be overwhelming, especially when using a traditional node-link visualization to see context around neighboring entities. Your task is to use visual analytics to help FishEye analysts see, interact with, and understand the context around a tip. Special emphasis should be placed on displaying relevant contextual information and hiding information that is not interesting or relevant. Ideally, the display will be dynamic and interactive. Using visual analytics, can you help FishEye identify companies that could be engaged in illegal fishing?

Tasks and Questions:

Use visual analytics to build a “contextualizer” that provides rich information for the entities listed below. Each entity was identified in a tip received by FishEye, but not all will end up with a substantive connection to illegal fishing. The visualizations you develop should allow analysts to extend the information they see as new entities are discovered.

Entities to investigate

  1. Mar de la Vida OJSC

  2. 979893388

  3. Oceanfront Oasis Inc Carrie

  4. 8327


  1. Use visual analytics to dynamically display and explore context around the suspected entities listed above. What did you learn about each one? Can you connect them to illegal fishing? Provide evidence for or against the case that each entity is involved in illegal fishing and use visual analytics to express confidence in your conclusions. Limit your response to 600 words and 6 images.

  2. Use your visual analytics tool to compare and contrast what you learned about the suspect entities. Are there patterns that may indicate illegal activity? Use visual analytics to express confidence that a pattern exists and where uncertainty may be affecting this confidence. Limit your response to 400 words and 4 images.

  3. What other companies should FishEye investigate for illegal fishing? Show how your visual analytics can be used to find entities that are worthy of further investigation. Limit your response to 600 words and 6 images.

  4. Reflection: What was the most difficult aspect of working with this knowledge graph? Did you have the tools and resources you needed to complete the challenge? What additional resources would have helped you? Limit your response to 300 words

Note: Participants in MC1 should use only data from MC1 for their submissions. Use of external data, including from other mini-challenges or external sources, is discouraged. Participants interested in combining data from other challenges are encouraged to participate in the Grand Challenge.

Clarification Requests

1. Questions on aspects of the data

Question There are some nodes which have no unique id identification and we get a number of nodes differing from the data notes file. The number of relationships also differs. What could be happening?


The graph does have some noise and inconsistencies, as observed in any real-world graph could cause issues related to the question. With regards to number of relationships, It is possible you are treating every duplicate multi-edge as a new edge. The graph JSON may have two exactly the same edges coming from two different documents. The number listed in the date document are based on unique edges. #### 2. QUestion on data encoding ####


We assumed the encoding was UTF-8, but there’s some characters that are not legible. How should we parse this?


utf-8 is the default encoding used in the MC1 graph. Since this graph is curated from real-world news sources, there are multiple instances of unresolved Unicode characters. It is suggested to use utf-8 encoding for graph loading and that will keep the graph structure and attributes consistent

VAST 2023 Submission Instructions