When people use the application, they're performing actions (viewing, creating, modifying, and so on) on objects (documents, blog posts, user profiles, and more). The application keeps track of these actions and events. The analytics feature takes that information and transforms it into data you can use to learn more about how people are using the application over time. The analytics data is stored according to a data model that's commonly used for other data warehousing and applications. Through this standard model, you can more easily retrieve the kind of data you need, analyzing it even in simple applications such as a spreadsheet program.
In addition to this real-time activity collection, there are also batch jobs that run at night, populating the data store with supplemental data. That data includes the names of content and users associated with these activity records. This process of batch loading is known as an "extract, transform, and load" (ETL) process.
The final result is a data store (sometimes referred to as a "data mart" or "data warehouse"). To this data store, you can connect other reporting tools in order to generate visual reports. You can also connect this store to an OLAP product in order to achieve sophisticated exploration and analysis of the data.