Twitter/X Sentiment Analysis

This project classifies tweets as Negative, Neutral, or Positive using Machine Learning.

📊 Dataset

• ~75K tweets
• After cleaning: ~57K tweets
• Columns: id, entity, sentiment, tweet

Dataset: Twitter Entity Sentiment Analysis

⚙️ Workflow

1. Clean and preprocess tweets (remove links, mentions, hashtags, punctuation).
2. Convert text into numbers using TF-IDF.
3. Train a Logistic Regression model.
4. Evaluate with accuracy, precision, recall, and F1-score.

📈 Results

• Accuracy: ~79%
• F1-scores:
  • Negative: 0.82
  • Neutral: 0.76
  • Positive: 0.79

Sentiment distribution

🔍 Example

tweet = "I love programming! #coding" prediction = model.predict(vectorizer.transform([tweet])) print(prediction) # Positive