Anomaly detection is the task of identifying anomalies or outliers by analysing existing trends or behaviours in a particular dataset. In the case where we are searching for anomalies in time series data, we can exploit seasonal and trend patterns. This project was an anomaly detection system that I worked on as part of an internship at Auto & General. It is used to identify anomalies in time-series data using (primarily) a Seasonal-Trend decomposition with LOESS (STL) algorithm.

• GitHub