Phebi Adds Data Viz Tools to its Voice Analysis

London-based voice technology firm Phebi has released three data visualization tools helping researchers understand respondent emotions in Insight and Detect products.

The first tool, Emotion Tracker (pictured, above), display maps the 'shape' and 'color' of emotions over the course of voiced answers. Users can listen to voices / read transcriptions while watching emotions change, zooming out for high-level views or in for near-real time analytics (Phebi measures emotion every three seconds).

The second, Audio Explorer adds value to the audio data many qualitative researchers already collect from interviews and focus groups, helping them find keywords or phrases wherever they occur across vast audio files; and sorting responses by emotion, making it fast and easy to scan surrounding text for context.

The third, Emotion Clouds (pictured, below), reveal the words respondents use most frequently by emotional state. With them, the firm says researchers can quickly identify which words 'pop' when respondents feel strongly, anxious, calm, happy or sad, opening up new levels of understanding.

Co-founder and CEO Mike Page comments: 'Market researchers are always looking for an edge - a deeper level of understanding - to help them better predict behaviour. Speech, with its nuanced variations in pitch, tone, rhythm and volume, can be incredibly revealing about people's emotions but, historically, it hasn't been possible to interpret the emotion in audio data at scale'.

Co-founder and CTO Konstantin Morjan says the company was built 'to bust the trove of emotional insights wide open'. He adds: 'Using a combination of advanced machine learning techniques, contextual language models and AI, all trained and tested in our labs, we've analyzed emotion in voice responses since our launch. Now, with our new visualization tools, researchers can see, play with, sort, dissect and analyze emotion data in new ways that advance their understanding, help them tell better stories and enable better decisions'.

Home page: www.phebi.ai .

All articles 2006-23 written and edited by Mel Crowther and/or Nick Thomas, 2024- by Nick Thomas, unless otherwise stated.