🚀 Brand Sentiment & Image Index Dashboard
A production-ready analytics application that converts unstructured social and news data into a unified Brand Image Index (0–100) — enabling real-time tracking of brand perception, sentiment dynamics, and market positioning.
🎯 Why This Matters
Most sentiment tools stop at “positive/negative.” This system goes further by combining sentiment + attention (volume) + trend (momentum) into a single decision-making metric.
👉 Built for marketing strategy, brand analytics, and investment insights.
🧠 Core Architecture
**Pipeline Design: ** Data Collection → NLP Processing → Index Engine → Dashboard Data Sources Reddit (PRAW API) Twitter/X (snscrape) News (RSS feeds) NLP Engine VADER + TextBlob (ensemble scoring) Keyword extraction for insights Extendable to transformer models 📊 Image Index Model Image Index = (0.5 × Sentiment) + (0.3 × Volume) + (0.2 × Momentum) Sentiment → Perception Volume → Awareness Momentum → Direction
✨ Key Features:
📌 Multi-brand comparison 📈 Sentiment & volume trend analysis ☁️ Word cloud (positive vs negative signals) 📊 Real-time Image Index scoring 🔍 Raw mention-level insights 🧪 Demo mode (runs without API keys)
📂 Project Structure
brand_sentiment_app/ ├── app.py # Streamlit dashboard ├── scraper.py # Data collection layer ├── sentiment.py # NLP pipeline ├── image_index.py # Scoring engine ├── requirements.txt └── Brand_Sentiment_Image_Index.ipynb
⚙️ Getting Started Run Locally git clone cd brand_sentiment_app pip install -r requirements.txt streamlit run app.py
** Run on Colab** Open notebook Run cells (auto setup) Access via ngrok link Deploy (Streamlit Cloud) Connect GitHub repo Select app.py Add optional API keys Deploy in ~2 minutes
🔑 Environment Variables (Optional)
REDDIT_CLIENT_ID=your_id REDDIT_CLIENT_SECRET=your_secret
🧩 Tech Stack
Python Streamlit PRAW snscrape feedparser VADER, TextBlob
📈 Use Cases Brand monitoring & reputation tracking Marketing performance analysis Competitive benchmarking Sentiment-driven investment signals
🔮 Roadmap Transformer-based sentiment (BERT) Real-time streaming pipeline Predictive brand scoring Financial market correlation models
🤝 Contribution
Open to improvements and extensions. Fork, build, and raise a PR.
📄 License
MIT License