A Streamlit-based web application that detects whether a given text contains hate speech, offensive language, or clean content (Not hate) using a Decision Tree Classifier trained on labeled tweet data.

This project demonstrates a simple yet effective Natural Language Processing (NLP) pipeline for classifying text as Hate Speech, Offensive Language, or Not Hate Speech. The app allows users to input any text, which is then cleaned, processed, and classified using a trained Decision Tree model.

It includes a visually appealing Streamlit interface with color-coded results for easy interpretation.

🔗 Want to try it out? Visit the live app here:
👉 https://hatespeechdetection-vikash.streamlit.app

• 🧼 Text Preprocessing: URL removal, punctuation cleaning, stopword filtering, and stemming.
• 💬 Real-Time Prediction: Classifies input text instantly when submitted.
• 🎨 Custom UI Styling: Styled using HTML and CSS for a clean and modern look.
• 📊 Model Training & Evaluation: Decision Tree Classifier trained using CountVectorizer features.
• 💾 Pickle Integration: Uses preprocessed data and trained model stored as .pkl files.

git clone https://github.com/yourusername/hate-speech-detection.git
cd hate-speech-detection

python -m venv venv
source venv/bin/activate       # On Linux/Mac
venv\Scripts\activate        # On Windows

Create a requirements.txt file with the following:

streamlit
nltk
scikit-learn
pandas
numpy
matplotlib
seaborn

Then install them:

pip install -r requirements.txt

python -m nltk.downloader stopwords

streamlit run app.py

• Vectorizer: CountVectorizer()
• Algorithm: DecisionTreeClassifier()
• Training Data: Preprocessed tweets labeled as
  • Hate Speech
  • Offensive Language
  • Not Hate Speech

The model is trained on cleaned text data (clean_data.pkl) and label data (dataset.pkl).

The cleaning() function performs several preprocessing steps:

1. Lowercasing
2. Removing URLs, HTML tags, mentions, digits, and punctuation
3. Removing stopwords
4. Stemming words using SnowballStemmer

• 🔍 Integrate advanced models like Logistic Regression, Random Forest, or BERT.
• 🗣️ Add multilingual support for hate speech detection.
• 📊 Include data visualization dashboards for model insights.