A Comprehensive Study of Capgemini Employee Reviews
This project aims to analyze Capgemini employee reviews to uncover trends and patterns in employee satisfaction. The objectives include identifying primary factors influencing satisfaction, developing an NLP model to extract keywords from reviews, and creating a recommendation system for tailored suggestions. Additionally, a multiclass sentiment classification model will be developed to categorize employee ratings as positive, negative, or neutral. By leveraging advanced data analysis and machine learning techniques, this project seeks to provide insights into employee sentiments and improve organizational strategies for enhancing employee satisfaction and engagement.
The Steps I Took
- Data Collection and Cleaning
- Exploratory Data Analysis (EDA)
- Natural Language Processing (NLP) Model Development
- Recommendation System Implementation
- Multiclass Sentiment Classification Model
Why I Took These Steps
- Ensured reliable data for analysis
- Identified key factors influencing satisfaction
- Extracted meaningful insights from feedback
- Enhanced employee experience and engagement
- Improved sentiment analysis accuracy
Tools I Used
-
Excel
-
Python
-
Power BI
Challenges:
- Missing Values: Certain columns contain irreplaceable missing values, which may impact the model's accuracy and reliability.
- NLP Model Limitations: NLP models may struggle with understanding context in employee reviews, affecting the accuracy of keyword extraction, sentiment analysis, and recommendations.
- Imbalanced Data: Data imbalance can negatively affect the performance of the sentiment analysis model. Techniques like oversampling or undersampling can introduce bias or lead to a loss of information.
- Generalization Issues: The multiclass sentiment classification model may struggle to generalize if the training data lacks diverse sentiments across various organizational contexts, which can impact its accuracy on new or unseen data.
