Thomson Reuters Web of Knowledge Build Challenge Winner Q&A

November 2013
Innovation Challenge Winner Balazs Godeny

In May 2013, Thomson Reuters launched the Build Challenge, a contest for the scientific and scholarly research community designed to expand the discovery experience offered by Thomson Reuters Web of KnowledgeSM, an intelligent research platform providing access to the world's leading citation databases with over 20 million users, encompassing students, information professionals, researchers and teaching faculty. The Build Challenge received 15 total submissions.

The Build Challenge is the second in a series of innovation challenges to be hosted by the Scientific & Scholarly Research group of Thomson Reuters. The program is designed to raise awareness of relevant news and commentary, recognize excellence in research, and encourage professional development. This challenge awarded $10,000 USD to the selected individual who developed a prototype with workable code and an innovative concept paper demonstrating new ways users can interact with and explore data relationships within Web of Science content.

 


 

Interview With Innovation Challenge Winner Balazs Godeny

What motivated you to participate in the Web of Knowledge Build Challenge?

The challenge raised several questions that I have always found interesting, like how to visualize relationships of complex objects, how to guide the users by combining their human intelligence with the machine's computing power, how to mine meaning from text documents and many more. When I read the task description in the InnoCentive newsletter I immediately knew that this could be a fun project to work on.

Can you tell us a bit more about your solution?

The application I created lets the user overview a set of documents — typically a result set of a search the user executed — together with their connections to other documents where the importance of each document and also the type and strength of the connections are visualized. Unlike a traditional search result list, the results are displayed as nodes on a map, where the nodes correspond to the most important keywords or authors found in the whole document set. Moreover, the distance between the nodes on the map can be interpreted as semantic relatedness of the nodes. If, for example, two keywords are displayed close to each other, then the keywords are likely to co-occur in many documents.

Innovation Challenge screenshot

How did you come up with the idea for this solution?

I was trying to create a system that displays as much meaningful information about a whole set of documents as possible, and of course doing it without overflowing the user with too much data. The main idea is that if we want to guide the users' exploration process then showing the presence, nature and strength of connections between documents may be much more important for them to see than the details of the individual documents.

How do you envision seeing your solution implemented in Web of Knowledge? In what way will this benefit users of Thomson Reuters Web of Knowledge? Do you have any suggestions for the Web of Knowledge product team on how best to build on your idea?

My solution is more than just an idea; rather, it is a working system that combines several new ideas that are not present in the current system. I hope that some of these new ideas could easily be incorporated into the Web of Knowledge platform and also that doing so can significantly improve the users' search experience. Based on the judges' decision of awarding me the first prize it seems they share this hope too, but as I don't have any knowledge of the current architecture I'd rather not suggest anything on how to do it in practice.

How will this solution affect your workflow and the work of other researchers?

The problem that I was hoping to solve is the frequent case in research when you are not exactly sure what you are looking for. Finding a needle in a haystack is not that difficult, provided you know that what you are searching for is a needle and you have the appropriate tool — for example, a very strong magnet. Finding documents in a library typically doesn't work this way. If you knew exactly what you wanted then any search engine would get you there easily. But most of the time you are not searching for a well-defined object, rather you are exploring, researching, investigating, following threads, observing connections and noticing similarities. At each step of this process you are making decisions that are based on your past experience, human intelligence and intuition. The application I created aims to be a tool that helps researchers make intelligent decisions that are based on solid facts. You add the human element, the system provides the data, presenting them in a way that helps the most in your exploratory process.

The data and citation records included in this report are from Thomson Reuters Web of ScienceTM. Web of ScienceTM is a registered trademark of Thomson Reuters. All rights reserved.