Which resources in Azure Machine Learning are used to explore data?

Data exploration in Azure Machine Learning primarily leverages data assets and data stores. Data assets represent datasets that are stored in various formats, making it easy for data scientists to access and analyze the data directly within the Azure Machine Learning workspace. These data assets can be linked to data stores, which are repositories that provide a connection to the actual data, whether it is stored in Azure Blob Storage, Azure SQL Database, or other services.

By utilizing data assets and data stores, data scientists can perform tasks such as data cleaning, transformation, and preliminary analysis, which are crucial for developing an effective machine learning model. This setup allows for an organized approach to managing the data lifecycle, ensuring that datasets can be versioned and reused efficiently.

In contrast, compute instances provide the processing power needed for training models or running experiments, but they are not primarily focused on data exploration. Web applications serve a different purpose, typically aimed at deploying models and providing user interfaces for interaction rather than data exploration. Virtual machines might also be used for various computing needs, but they are not specific to Azure Machine Learning and do not inherently provide the structured framework for data exploration as data assets and data stores do.