Data Cloud Video Hub

2. Data prep for features and labels

Einstein Studio continues its exploration of predictive AI in Data 360 focusing on data preparation for loan application model's features. The session emphasizes the importance of transforming complex data sets into a denormalized format, allowing for effective model training with a single row per application ID. It introduces tools such as transforms and calculated insights for batch and streaming feature preparation, while highlighting the need to avoid data leakage by ensuring that future information isn't included in training data. The discussion also covers scheduling updates and the optimal use cases for each tool to enhance model performance.