Mining and Visualizing Online Web Content Using BAM: Brand Association Map

Authors

  • Navot Akiva Nielsen Online Inc
  • Eliyahu Greitzer Nielsen Online Inc.
  • Yakir Krichman Nielsen Online Inc
  • Jonathan Schler

Abstract

In this paper, we describe our Brand Association Map (BAM) tool which maps and visualizes the way consumers naturally think and talk about brands across billions of unaided conversations online. BAM is a semi-supervised tool that leverages text-mining algorithms to identify key correlated phrases, terms and issues out of millions of candidate terms which were derived from billions of online conversations. The most correlated phrases with a given brand are then projected and plotted onto visual bull's eye representation. BAM's visualization illustrates both the correlation level between a brand (appears in the center of the visualization) and each of the highly correlated terms as well as the inner correlations among all presented terms, where terms on the same radial angel represent a "clustered" discussion of terms frequently mentioned together. We found BAM useful for extracting various intuitions and beliefs that are highly correlated with brands to better grasp how consumers really contextualize them, out of massive consumer generated media (CGM) documents.

Downloads

Published

2021-09-25

How to Cite

Akiva, N., Greitzer, E., Krichman, Y., & Schler, J. (2021). Mining and Visualizing Online Web Content Using BAM: Brand Association Map. Proceedings of the International AAAI Conference on Web and Social Media, 2(1), 170-171. Retrieved from https://ojs.aaai.org/index.php/ICWSM/article/view/18635