team_comm_tools: A Python Toolkit for Exploring the Communication Space

How a team talks is a rich source of information about its collaboration process, revealing how its members resolve conflicts, coordinate activities, and monitor goal progress. However, the full potential of communication data often goes unrealized: analyzing communication requires a series of contingent decisions about which constructs to study and how to measure them. Consequently, researchers interested in team communication face a high barrier to entry, challenges with reproducibility, and constraints on their ability to capture phenomena that emerge from interactions between multiple conversational dimensions. To address these challenges, we present the Team Communication Toolkit (team_comm_tools on the Python Package Index), an open-source Python package that extracts more than 160 communication features from any corpus of text-based conversations. Our collection of measures allows the analyst to understand the effect of conversational features in the context of a broader family of features that may equivalently explain the phenomenon. Rather than forcing researchers to commit to a single measurement of complex phenomena such as constructiveness or assertiveness, the toolkit’s flexible output format makes it possible to systematically compare measures, isolate confounders, and identify higher-order interactions. We call our general analytical approach the “Communication Space,” reflecting the fact that researchers can now more easily explore the realm of possible ways to analyze conversational data. Finally, we demonstrate the utility of our tool and framework through a case study exploring conflict escalation in both real-world and synthetic data.