Network and Sentiment Analysis of Enron Emails

TL;DR

This study analyzes Enron's email network and sentiment to identify key employees, community structures, and information flow, revealing insights into organizational dynamics and sentiment trends around the company's crisis.

Contribution

It combines network analysis and sentiment analysis on Enron emails, examining threshold effects and their impact on community detection and importance measures.

Findings

Community structures aligned with formal organization.

Importance of individuals varies with centrality measures.

Email sentiment did not predict financial crisis.

Abstract

The objective of the research was to analyze e-mails exchanged at Enron, a power company that declared bankruptcy in 2001 following an investigation into unethical operations regarding their financials. Like other researchers, we identify the most important employees and detect communities using network science methods. We find that the importance of a person depends on the centrality measure used; while the communities we detected resembled the formal organizational structure of the company. In addition, because previous work required that 10 e-mails be sent and received for an e-mail relationship to exist, we analyzed the effect of different thresholds on the results and found that results were very dependent on the threshold used. We also performed sentiment analyses on the e-mails to evaluate whether sentiment changed over time and found that the sentiments of the e-mails do not give insight into the financial wellbeing of Enron. Our results provide insight into how information flowed through Enron, who the key employees were, and e-mail sentiment before and after the crisis