As dashboards enable you to quickly add and remove filters to perform data analysis, they require all data to be present in memory (specifically, a results section) for the filters to work. If the dataset is small, it is not a problem, but a big dataset can overwhelm the memory capacity of a desktop computer. You may not want to wait for a big dataset to be loaded and processed, if you are concerned only with a section of the data.
To solve this problem, you can apply filters to the data query, or underlying data model. See Drill Down on Limits and Filters. The query is processed, and the relevant data is retrieved from the database, reducing the time and memory needed to perform the analysis. The Tree component enables you to quickly apply a hierarchy of filters to a query.