Challenge
Our client, a large-scale automotive components manufacturer, had recently implemented a real-time health feedback solution for fleets of vehicles using their components. The system used IoT engine and sub-system data points to create real-time notifications of issues with the vehicles along with recommendations for correcting those issues. This system was implemented in Azure using Event Hub functionality and custom development.
While this solution was successful, it was also limited to only reactively responding to warnings and errors as they occurred. The company needed an analytics solution to create functionality for developing predictive models that provide meaningful proactive recommendations to the fleets.
Solution
To support the high-level analysis and machine learning required to build predictive models, a large amount of data over an extended time frame was needed. Since the current system only retained data long enough to provide the real-time responses, another big data solution was required to capture and retain the IoT data.
Additionally, there was a need to integrate manufacturing data with the IoT data to be used in the analytics processes. This was accomplished by designing and implementing a Data Warehouse to collect and organize the data and provide a data platform for high-end analytics and machine learning.
Ultimately, the solution uses the existing Event Hub functionality to output the data, Azure Data Factory to manage the data flow, and Snowflake Data Warehouse to store, organize, and integrate the data.
