Anomaly Detection Platform
With the help of Anomaly Detection (AI-Form), data streams with learned patterns are continuously examined for deviations. The user is informed when a change is reliably detected.
A major German telecommunications provider offers an anomaly detection service for its end customers based on its IoT platform.
Our tasks were/are the design of the architecture and usability of the application. The software development of the components. The operation of the application based on the IoT platform.
The challenge was to implement a generic application that could be set up and used independently by the end customer without any knowledge of Data Analytics / Data Engineering.
The solution was a guided wizard that supports the end customer during the setup process. Based on the selected machine and sensor data, this wizard provides recommendations as to which ETL/ ELT process can be used for the transformation and which algorithm can be used to detect the anomalies. The system learns based on the anomalies confirmed by the end customer.