MAPINTEL – Interactive visual analytics platform for competitive intelligence, 2019
MAPINTEL – Interactive visual analytics platform
for competitive intelligence, 2019
The ability to quickly identify, process and understand the impact of political, economic and social events in the world arena constitutes a fundamental requirement to accomplish AICEP’s mission. In an activity in which timing and assertive responses are of the essence, having analysts sorting through the Internet to find relevant news and updates is ineffective, time-consuming and expensive. In this project we propose a neural network-based document clustering system, which is a genuine content addressable memory system, meaning it clusters and stores text documents so they can be retrieved based on their content. Using Self-Organizing Maps (SOM’s), documents are mapped as points on a semantic map, in a topological order that describes the similarity of their contents.
This map is then used as a metaphor to interact with the text corpus, creating an interactive visual analytics tool, allowing for its exploration through different, but correlated views and iteratively selecting and examining documents.