For many newcomers to Azure Cosmos DB, the learning process starts with how to model and partition data effectively. How should I think about modeling data in Cosmos DB? When should I co-locate data in single collection verses multiple collections? When should I de-normalize or normalize properties in the same document vs multiple documents? How should I apply a partition key to this object model? In this session, we discuss the strategies and thought process one should adopt for modeling and partition data effectively in Azure Cosmos DB. We also briefly cover related topics such implementing optimistic concurrency control, transactions with stored procedures, batch operations, and tuning queries + indexing.