Skip to content

RyanDing2316/LLM-Training---Final-Assessment-Practice-Task

Repository files navigation

Employee Sentiment Analysis — README

Summary

This project analyzes employee email messages to evaluate sentiment and engagement across time. The pipeline includes sentiment labeling, exploratory data analysis, employee scoring and ranking, flight risk detection, and future sentiment prediction.


Top 3 Positive and Negative Employees (Latest Month)

Top Positive Employees:

  1. lydia.delgado@enron.com
  2. johnny.palmer@enron.com
  3. sally.beck@enron.com

Top Negative Employees:

  1. rhonda.denton@enron.com
  2. kayne.coulter@enron.com
  3. don.baughman@enron.com

(Note: Rankings are based on sentiment scores in the most recent available month.)


Flight Risk Employees

Employees who sent 4 or more negative emails within a 30-day period were flagged as flight risks. Each employee is flagged once per distinct 30-day window (no overlapping flags).

Flagged Flight Risk Employees:


Key Insights

  • Sentiment Distribution: Negative and neutral messages dominate overall communications.
  • Employee Trends: Some employees consistently show high or low sentiment across months.
  • Flight Risk Timeline: Repeated detections show chronic disengagement in specific employees.
  • Time Trend: Sentiment scores are increasing slightly over time, indicating potential improvement.

Recommendations

  • Prioritize employee feedback and engagement efforts for those repeatedly flagged as flight risks.
  • Investigate root causes of negative sentiment among lowest-ranked individuals.
  • Consider using this pipeline as part of a real-time HR monitoring system.
  • Extend analysis with advanced models or integrate department/team data for deeper insights.

*For full details, see the project report and associated visualizations in the visualization/ folder.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors