System Architecture & Design Flashcards

Which scenario best justifies exporting raw analytics data to a data warehouse such as BigQuery?

Answer: Joining event-level analytics with CRM data for custom analysis

Raw exports enable event-level joins with external sources and custom modeling beyond UI reports.

What does ETL stand for in analytics data engineering?

Answer: Extract, Transform, Load

ETL describes extracting data from sources, transforming it, and loading it into a target store.

Data sampling in an analytics tool occurs mainly because:

Answer: Queries over very large datasets are limited to a subset to return results faster

Sampling trades precision for speed when processing exceeds a data threshold.

Which design choice most reduces the risk of breaking reports when the website's HTML is redesigned?

Answer: Using data-layer pushes for key events

Data-layer events are independent of markup, so redesigns don't break tracking.

A single-page application (SPA) loads new content without a full page reload. What must be implemented for accurate pageview tracking?

Answer: Virtual pageview or history-change triggers

Because the page doesn't reload, tracking must be triggered by route or history changes.

Which component of an architecture is responsible for removing internal traffic and known bots before data reaches reports?

Answer: Filtering/processing layer

Filters applied during processing exclude unwanted traffic from reporting data.

Why should a staging environment use a separate analytics property or data stream from production?

Answer: To keep test data from polluting production reports

Separate properties keep QA traffic from contaminating real metrics.