senior design
DeNOVO: Pattern of Life
Anomaly detection for time series data.
DeNOVO: Pattern of Life finds abnormal patterns in time series data. It was the CU Denver senior design project for Fall 2024 and Spring 2025.
Someone uploads a time series dataset, trains anomaly detection models, reviews the results in interactive charts, and exports the findings.
Detection
The models are Isolation Forest, DBSCAN, and polynomial regression. They cover single odd points, groups of related points, and values that leave a fitted trend.
What I worked on
Backend architecture, the Isolation Forest path, preprocessing for mixed time series files, and the database schema. The team was Kye Dombrowski, Gaurab Baral, and Jacob Moreno, with Debra Parcheta as mentor.
The app is containerized with Docker. CI ran on Bitbucket Pipelines. There is no public demo.