This project explores anomaly detection techniques for identifying unusual patterns in data.
- Time series feature engineering to capture trends and variability
- Anomaly detection using clustering-based methods with Silhouette score evaluation
- Dimensionality reduction (PCA) to improve detection performance and interpretability
- Clever data visualizations to overlay detected anomalies on original time series
- Interactive dashboards built using Dash for dynamic exploration of results
- Python
- pandas
- numpy
- scikit-learn
- matplotlib
- Dash