LSU researchers create topic map of oil spill disaster
POSTED: Tuesday, July 17, 2012 - 6:00am
UPDATED: Wednesday, October 24, 2012 - 9:07am
Baton Rouge, La — The Gulf of Mexico Deepwater Horizon Oil Spill incident has impacted many aspects of the coastal environment and inhabitants of surrounding states. However, government officials, Gulf-based researchers, journalists and members of the general public who want a big picture of the impact on local ecosystems and communities are currently limited by discipline-specific and fractured information on the various aspects of the incident and its impacts.
To solve this problem, Assistant Professor in the School of Library and Information Science Yejun Wu is leading the way in information convergence on oil spill events. Wu’s lab has created a first edition of an online topic map, available at http://topicmap.lsu.edu/, that brings together information from a wide range of research fields including biological science, chemistry, coastal and environmental science, engineering, political science, mass communication studies and many other disciplines in order to promote collaboration and big picture understanding of technological disasters.
“Researchers, journalists, politicians and even school teachers wanted to know the impacts of the Deepwater Horizon oil spill incident,” Wu said. “I felt this was an opportunity to develop a tool for supporting learning and knowledge discovery. Our topic map tool can help people learn from historical events to better prepare for the future.”
Wu started the project with a firm belief in the need for an oil spill information hub.
“There is a whole list of historical oil spill events that we probably neglected – we did not learn enough from history,” Wu said.
He first looked to domain experts from various disciplines to share their own views of the impacts of the Deepwater Horizon oil spill. From there, Wu and his research associate and graduate students manually collected more than 7,000 concepts and 4,000 concept associations related to oil spill incidents worldwide from peer-reviewed journal articles and authoritative government websites, loading the information into an organizational topic map software program. Prior to these efforts by Wu’s lab, no comprehensive oil spill topic map or taxonomy existed.
“Domain experts typically focus on oil spill research in their own area, such as chemistry or political communication, but an oil spill is a comprehensive problem, and studies should be interdisciplinary,” Wu said. “Experts in different fields that usually don’t talk to each can benefit from a tool that brings together and organizes information concepts across many disciplines.”
As many graduate students are perhaps well aware, writing an academic thesis on a scientific topic often means starting from scratch, scouring the peer-reviewed literature for concepts and relationships associated with a particular topic of interest. Topic maps are a new semantic approach to information organization designed to enhance such information navigation and retrieval from the outset.
“Without a previous knowledge structure, it’s hard to know what’s going on,” Wu said.
The topic map created by Wu’s lab provides a roadmap of oil spill impacts worldwide, pulling information and references together in one spot to enhance understanding and research related to oil spill impacts. Wu expects the map to facilitate both understanding of specific impacts as well as knowledge discovery through interdisciplinary knowledge fusion.
The topic map is composed of three major levels: concepts or topics, associations or relationships, and the information sources for these concepts and associations.
“After the BP oil spill, chemical dispersants impacted shrimp populations in certain areas,” Wu said. “In this case, ‘chemical dispersants’ and ‘shrimp’ are the topics while the verb ‘impact’ is the association. We can also trace this statement back to its information source within the topic map.”
The topic map is both visual and text-based, linking together concepts such as “oil rig explosion,” “coastal business” and “threatened and endangered species” in a visual relationship web. This web of information can help scientific researchers and lay audiences alike make connections between causes and effects related to oil spill incidents. For example, Wu’s topic map establishes a “cause” link between the concepts of “Deepwater Horizon 2010 in Gulf of Mexico” and “anger,” as well as an “influenced by” link between “anger” and “likelihood of being politically active and engaged.” The topic map can thus help oil spill researchers in non-communication fields understand spill impacts on public political engagement, a link that might not be readily identified by these researchers otherwise.
The topic map itself is a feat of interdisciplinary collaboration, the product of hundreds of man hours and collaboration between Wu and graduate students spanning the fields of computer science, chemistry, and English studies.
“This is a large, interdisciplinary tool for discovering concepts and associated areas of oil spill research,” Wu said.
Wu’s topic map also caters to the varying information needs of different users, with a ‘tourism’ visual map of interesting points to see geared toward lay audiences as well as an underlying text version of the topic map showcasing a more complete set of concepts and associations for expert researchers. Oil spill researchers may find the map useful for interdisciplinary research and knowledge discovery purposes, while teachers may find the map useful in teaching their students about oil spill incidents and their wide impacts to coastal environments. Journalists may find the map useful in creating stories that incorporate big picture perspectives of oil spill incidents, while government officials may find the map helpful in forming policies and planning for future disasters.
Wu’s lab is currently undertaking automatic information extraction from more than 700 documents, looking to incorporate more than 10,000 concepts into the topic map. Wu is also testing his topic map for usefulness and ease of use with high school students.
“In the next phase of testing, we will be recruiting college students and professors to play with the tool to see whether our map can serve its purpose of facilitating knowledge discovery and interdisciplinary learning,” Wu said. “I believe this tool can be useful in many different aspects, some perhaps beyond my current knowledge.”