May 16, 2024

Describe clusters

Our topics analysis tool has been enhanced with a new feature called "Describe Clusters." This feature significantly improves the way we group and name clusters of free text based on their similarity and the topics they cover.

Previous Method:

  • Previously, clusters were named by combining significant terms from each cluster. While this method was functional, it often resulted in names that were difficult to read and did not effectively communicate the insights.

New Improvement:

  • With the integration of OpenAI models, we now generate titles for clusters that are much more readable and contextually accurate. This enhancement takes into account the context of all paragraphs, ensuring that the titles are not only more human-readable but also more meaningful.

Example:

  • When analyzing texts related to product management, if a cluster frequently mentions the word "design," the new feature will likely identify it as "UX/UI Design," aligning it more closely with the field of product management rather than general design.

Applications:

  • This feature is highly beneficial for organizing and understanding reviews, transcripts, emails, survey answers, book paragraphs, and any other free text. It provides clear and concise cluster titles that enhance the overall analysis and communication of insights.

We are excited to bring you this update, which aims to make your text analysis experience smoother and more insightful.