The Highlights feature is an overview in understanding the contents of your text analysis project. In this feature, orient yourself and start navigating the primary questions you can answer:
Theme detection — What are people talking about?
Differential analysis — What differences appear across your metadata groups?
Score drivers — What is driving your Customer Experience up or down?
Emotional analysis — What are the overall and specific sentiments?
You’ll first view Journey Insights Hub’s answers to these questions in Highlights, and then engage deeply with each data perspective as you move through the Volume, Galaxy, Drivers and Sentiment features.
Understand the Highlights previews
Volume – Which concepts are most prevalent?
Volume – Which concepts are most prevalent?
A concept is a meaningful word, phrase, or topic identified in your data. Journey Insights Hub determines prevalence by considering how often a concept appears in the text analysis project compared with how commonly it appears in the language more generally.
A concept can appear in many records without being particularly distinctive, while a less frequently mentioned concept can still be highly prevalent because it is unusually specific to the text analysis project.
Each bar includes:
Exact matches: Records that contain the concept itself, including basic word-form variations
Conceptual matches: Records that express a closely related meaning without necessarily using the exact term
Total matches: The number of records containing either an exact or conceptual match
A concept is counted once per matching record. Repeating the same concept several times in one record does not create additional matches.
Use Volume when you want to compare more concepts, examine the balance between exact and conceptual matches, or see how concept volume changes across time periods, categories, or scores.
Galaxy – What are the largest clusters of conversation?
Galaxy – What are the largest clusters of conversation?
Journey Insights Hub groups strongly related concepts into clusters based on how they are used throughout the data. Each colored group represents a different area of conversation, while the concepts within it provide an indication of what that cluster is about.
Clusters can help you:
Recognize broad themes without defining them in advance
Discover language people use around a familiar subject
See where one topic may connect with a larger conversation
Identify areas that may deserve their own concept list or deeper analysis
The colors make different clusters easier to distinguish; they do not indicate positive or negative sentiment. Use Galaxy to explore the relationships within and between these clusters. There, you can select concepts, examine related language and interact with the relationship map.
Drivers – What issues are affecting
Drivers – What issues are affecting
Use the dropdown to select an available numeric field, such as a satisfaction score, rating, NPS, or another numeric field in your metadata.
For each concept shown, the card compares:
The average score among records matching that concept
The overall average score across the text analysis project
A concept with an average meaningfully above or below the text analysis project average may be worth investigating. However, this indicates an association, not proof that the concept caused the score to change.
If the text analysis project does not include a numeric field, Drivers' results may not be available.
Open Drivers to compare frequency with score differences, examine concepts above or below the text analysis project average and investigate which topics may be connected with an outcome.
Sentiment – What do people feel strongly about?
Sentiment – What do people feel strongly about?
The Sentiment preview shows whether the language surrounding a concept is generally positive, negative, or neutral.
The percentages apply only to the records or matches associated with each concept. For example, a concept shown as 80% negative does not mean that 80% of everyone in the text analysis project feels negatively about it. It means that 80% of the qualifying matches for that concept were classified as negative. Sentiment is determined from the surrounding context, not from whether an individual word appears positive or negative.
Use Sentiment to examine a wider set of concepts, review the complete positive, negative, and neutral distribution and read the records behind the results.
AI Summary – What are the main takeaways?
AI Summary – What are the main takeaways?
If AI features are enabled for your text analysis project, your Highlights overview may also include an AI-generated summary. The summary uses the project data to provide a high-level overview of important themes, key points, and potential areas to investigate.
Use the summary to:
Get oriented when exploring a text analysis project
Identify themes and potential areas for further investigation
Generate questions for deeper analysis
Compare the written summary with patterns in the visualizations