Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Cloud Digital Leader practice bank has an additional 3,070 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
An e-commerce retailer currently uses Looker and BigQuery analytic queries to review past sales performance and understand why revenue fluctuated across different regions last quarter. Leadership now wants to move beyond retrospective reporting to meet two new objectives:
Which approach should the company take to transition from descriptive analytics to meeting these new objectives?
Generative AI is a specialized category of artificial intelligence that uses large foundation models to create new content—such as text, images, audio, video, and code—or to summarize and extract insights from unstructured data. It excels at natural language processing, conversational search, content creation, and document understanding by learning the underlying style and structure of massive training datasets.
The company wanted to automatically condense thousands of written customer support tickets into key themes and create personalized promotional graphics for an upcoming marketing campaign.
In enterprise architectures, traditional predictive AI and generative AI are often combined. For instance, a traditional ML model first analyzes historical tabular data to predict which customers are likely to churn, and generative AI then creates personalized retention messages tailored to each customer segment.
BigQuery is Google Cloud's serverless, highly scalable enterprise data warehouse. Running standard SQL analytic queries in BigQuery allows data analysts to aggregate, filter, join, and examine petabytes of historical data rapidly. This process forms the foundation of descriptive and diagnostic data analytics, providing organizations with deep historical context about their business operations.
SUM, AVG, and GROUP BY). They do not train a machine learning model to uncover hidden correlations and predict future outcomes for individual customers.The company needed to analyze five years of historical transaction logs to determine which product categories generated the most revenue during past holiday seasons.
Although standard BigQuery analytics is retrospective, organizations can transition to predictive modeling directly within the same platform by using BigQuery ML. Instead of running standard summary queries, analysts use SQL statements like CREATE MODEL and ML.PREDICT to train and deploy machine learning models on their existing warehouse data.
Predictive machine learning (ML) is a branch of artificial intelligence (AI) that uses mathematical algorithms to automatically learn patterns and relationships from existing historical datasets. Unlike traditional data analytics and business intelligence (BI), which focus on retrospective analysis to describe what happened in the past and why, machine learning builds predictive models capable of forecasting future outcomes and estimating probabilities for unseen events.
While descriptive analytics and BI dashboards provide essential historical context by answering what happened and why it happened, they cannot extrapolate complex patterns to estimate individual future probabilities. Training predictive machine learning models on historical data is the optimal choice because it directly addresses the need to forecast continuous revenue figures and calculate future churn risk.
Looker is an enterprise business intelligence (BI) and data visualization platform that connects to governed data repositories like BigQuery. It enables organizations to access, analyze, and visualize up-to-date data through interactive dashboards, charts, and self-service reports. Even when configured for real-time monitoring, BI tools perform descriptive and diagnostic analytics—summarizing events that have already occurred or are happening at that exact moment.
The scenario asked how the company could provide executives with an up-to-date, governed visual summary of current subscription cancellation rates and regional sales performance.
A common misconception is that making business intelligence faster (moving from quarterly reports to real-time dashboards) makes it predictive. While Looker can display predictions generated by BigQuery ML or Vertex AI, a standalone BI dashboard monitoring current events remains descriptive analytics, not predictive modeling.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated