To avoid costly product recalls and decreased customer satisfaction, the Bosch data science team needed to predict when parts in production would fail. In addition, they needed more effective communications and collaboration across the company with product stakeholders and other data scientists who create data models.

The team had already developed predictive maintenance models, but needed a way to manage and seamlessly push them into production. Previously, models were created by teams around the world, with no way of searching and auditing past work. Because different data types are needed to predict part failures, the company also needed a way to connect to and blend data from databases and Hadoop environments.

TIBCO® Data Science enabled Bosch's data scientists to implement models in a fully managed and central environment. The TIBCO Data Science SDK was used to optimize data transformation workflows to quickly process large amounts of complex manufacturing data. TIBCO Data Science project management features made it easy to track the progress of models being created and pushed into production as well as to bring the right stakeholders from across the business into the process.