Quantitative analysis of communication dynamics in agile software teams through multimodal analytics

Area of Science:

• Computer-Supported Cooperative Work (CSCW)
• Human-Computer Interaction (HCI)
• Software Engineering

Background:

• Effective agile software development relies on communication, cooperation, and coordination.
• Coordination, crucial for collaborative tasks, is challenging to measure objectively.
• Limited empirical data exists on the impact of agile coordination techniques on team communication dynamics.

Purpose of the Study:

• To investigate the effect of coordination techniques on team communication patterns.
• To evaluate the utility of Multimodal Analytics (MmA) in analyzing collaborative interactions.
• To provide empirical evidence on how coordination influences speaking and attention distribution.

Main Methods:

• A controlled experiment involving 72 undergraduate students in 18 teams of four.
• Teams performed software effort estimation tasks with and without planning poker (a coordination technique).
• Audiovisual recordings were analyzed using MmA to measure speaking time (speech diarization) and attention time (facial orientation).

Main Results:

• Planning poker did not significantly change total speaking or attention time.
• Coordination techniques led to a more equitable distribution of speaking time among participants.
• MmA effectively captured subtle shifts in team communication dynamics.

Conclusions:

• Coordination techniques, such as planning poker, can enhance balanced participation in agile teams.
• Multimodal Analytics (MmA) is a valuable tool for objectively assessing collaborative communication.
• This study contributes empirical evidence supporting the role of coordination in improving agile software development collaboration.