Technologies Used
Understanding user sentiment is crucial in areas like customer feedback analysis, social media monitoring, and product reviews. I developed an interactive web application that performs real-time sentiment analysis using Python, Streamlit, and the VADER sentiment analysis model.
The application allows users to input text dynamically through a clean sidebar interface and instantly receive a detailed breakdown of sentiment scores — positive, negative, neutral, and compound. The compound score is then interpreted to determine the overall sentiment classification.
To improve usability and presentation, I implemented custom CSS styling within Streamlit, giving the application a clean, structured, and visually intuitive interface. The result is a lightweight yet powerful sentiment analysis tool that demonstrates both NLP integration and frontend customization using Python.
Key Features
- Real-time sentiment analysis of user-provided text
- Interactive and responsive Streamlit UI
- Sentiment scoring breakdown (positive, negative, neutral, compound)
- Automated overall sentiment classification
- Custom CSS styling for enhanced user experience
- Lightweight and easy-to-deploy architecture
Core Sentiment Analysis Logic
from vaderSentiment.vaderSentiment import SentimentIntensityAnalyzer
analyzer = SentimentIntensityAnalyzer()
sentiment_dict = analyzer.polarity_scores(user_input)
This integrates VADER’s pre-trained rule-based sentiment model to calculate polarity scores instantly.
Sentiment Interpretation Logic
if sentiment_dict['compound'] >= 0.05:
st.markdown("<h4 class='positive'>Overall Sentiment: Positive</h4>", unsafe_allow_html=True)
elif sentiment_dict['compound'] <= -0.05:
st.markdown("<h4 class='negative'>Overall Sentiment: Negative</h4>", unsafe_allow_html=True)
else:
st.markdown("<h4 class='neutral'>Overall Sentiment: Neutral</h4>", unsafe_allow_html=True)
The compound score is evaluated against defined thresholds to determine the final sentiment category.