3rd YearIntermediateAI/ML 4-5 weeks
Twitter Sentiment Analysis Dashboard
A pipeline that collects tweets on a topic, classifies sentiment with an NLP model, and visualises trends and keyword clouds on a live dashboard.
PythonNLTK / Hugging FaceTweepy / datasetStreamlitpandas
Find GitHub repos & source code →How to build it — step by step
- 1Ingestion: Collect tweets via the API (or a labelled dataset for offline use) and clean text.
- 2Classification: Use a pretrained transformer or train a classifier on a labelled corpus.
- 3Aggregation: Aggregate sentiment over time and by keyword; extract trending terms.
- 4Dashboard: Build a Streamlit dashboard with sentiment timeline, distribution, and word cloud.
Key features to implement
- ✓Text cleaning pipeline
- ✓Pretrained or custom classifier
- ✓Sentiment-over-time chart
- ✓Keyword word cloud
- ✓Topic comparison
💡 Unique twist to stand out
Add sarcasm/negation handling experiments and report how much they change accuracy versus a naive bag-of-words baseline.
🎓 What you'll learn
NLP preprocessing, text classification, transformers, and building data dashboards.