Sentiment Analysis in Python

This project demonstrates sentiment analysis in Python using two different techniques:

1. VADER (Valence Aware Dictionary and sEntiment Reasoner) - Bag of words approach
2. Roberta Pretrained Model from Huggingface Pipeline

Steps Covered in the Notebook:

Step 0: Read in Data and NLTK Basics

• Reading data using pandas
• Basic NLTK operations like tokenization and part-of-speech tagging

Step 1: VADER Sentiment Scoring

• Using NLTK's SentimentIntensityAnalyzer to get sentiment scores
• Applying VADER on the entire dataset
• Visualizing VADER results

Step 2: Roberta Pretrained Model

• Utilizing a pretrained model from Huggingface Transformers
• Comparing scores between VADER and Roberta model
• Visualizing scores between models

Step 3: Combine and Compare

• Pairplot to compare VADER and Roberta scores

Step 4: Review Examples

• Exploring examples where model scoring and review score differ
• Showing examples of positive 1-star and negative 5-star reviews

Extra: The Transformers Pipeline

• Demonstrating the Transformers Pipeline for quick sentiment predictions

How to Run

1. Ensure all dependencies are installed (pandas, numpy, matplotlib, seaborn, nltk, transformers)
2. Download or clone the notebook and associated dataset
3. Run each cell in the notebook sequentially

Dependencies

• pandas
• numpy
• matplotlib
• seaborn
• nltk
• transformers

Note

Make sure to replace tutorial_link_here in the README with the actual link to the tutorial.