Group Conversation Analytics using Raspberry Pi

Edtech

Python

Web

Signal Processing

IoT

Project overview

Developed during my first-year PhD research, this system captures and visualizes speaking dynamics in face-to-face group conversations. The interactive dashboard provides participants with real-time feedback about their conversation patterns, enabling more balanced and productive discussions.

Technical Implementation

The solution combines a Raspberry Pi with a ReSpeaker microphone array to create an edge-computing prototype that processes vocal interactions in real time. The system captures two key dimensions of group communication:

Vocal Activity Detection

Analyzes voice energy signals to identify speaking segments

Spatial Audio Analysis

Calculates direction-of-arrival (DoA) using beamforming techniques
Maps speakers to physical positions around the discussion table
Tracks speaking turns and participant engagement

Key Contributions

Developed a Python library for processing timestamped directional audio data.
Built an interactive dashboard (web-based) to visualize speaking patterns.

Skills Applied

Python, Raspberry Pi, Signal Processing, Web Development, IoT Prototyping

Libraries Used

pandas,matplotlib,networkx,numpy, dash