Study guide
2.3 Making Data Useful and Accessible
Google Cloud Digital Leader Exam
Track your progress as you read
Finished sections get a check mark.
1 Introduction & Exam Overview
1.1 Current Exam Facts 1.2 How to Prepare & Use This Guide 1.3 What Changed in the August 12, 2026 Exam Update
2 Digital Transformation with Google Cloud
1.1 Why Cloud Technology Is Transforming Business 1.2 Fundamental Cloud Concepts 1.3 Cloud Computing Models and Shared Responsibility 1.4 Agentic AI, Open Standards, and Google Cloud's Differentiators
3 Exploring Data Transformation with Google Cloud
2.1 The Value of Data 2.2 Google Cloud Data Management Solutions 2.3 Making Data Useful and Accessible 2.4 Data Types, the Data Supply Chain, and Interoperability
4 Innovating with Google Cloud Artificial Intelligence
3.1 AI and ML Fundamentals 3.2 Google Cloud's AI and ML Solutions 3.3 Building and Using Google Cloud AI and ML Solutions 3.4 Gemini, the Agent Platform, and the AI Hypercomputer
5 Modernize Infrastructure and Applications with Google Cloud — Part 1
4.1 Cloud Modernization and Migration 4.2 Computing in the Cloud 4.3 Serverless Computing
6 Modernize Infrastructure and Applications with Google Cloud — Part 2
4.4 Containers in the Cloud 4.5 The Value of APIs 4.6 Hybrid and Multi-Cloud
7 Trust and Security with Google Cloud
5.1 Trust and Security in the Cloud 5.2 Google's Trusted Infrastructure 5.3 Google Cloud's Trust Principles and Compliance 5.4 Google Threat Intelligence, Security Command Center, and Google SecOps 5.5 Securing the AI Stack: AI Protection, Model Armor, and Zero Trust
8 Scaling with Google Cloud Operations
6.1 Financial Governance and Managing Cloud Costs 6.2 Operational Excellence and Reliability at Scale 6.3 Sustainability with Google Cloud 6.4 Cloud Observability, SLIs, SLOs, and the Four Golden Signals
2.3 Making Data Useful and Accessible
Key Takeaways Explain This Topic AI
Looker is Google Cloud's business intelligence platform that democratizes data access through governed, model-based self-service analytics and dashboards built on top of BigQuery and other sources
Real-time streaming analytics creates business value that batch cannot, including fraud detection, IoT monitoring, personalization, and live operations, because decisions are made while events are still relevant
Pub/Sub and Dataflow are the core Google Cloud products for modernizing data pipelines: Pub/Sub ingests and decouples event streams, Dataflow processes them in stream or batch using Apache Beam
A modern analytics flow chains Pub/Sub for ingestion, Dataflow for processing, BigQuery for storage and analytics, and Looker for dashboards and embedded analytics
Last updated: August 2026
How Looker Democratizes Data Access
Quick Answer: Looker is Google Cloud's business intelligence (BI) platform that lets every team build trusted reports and dashboards on governed data, instead of waiting on a central analytics team. It models data once, in code, and exposes it to the whole organization through a browser.
Looker's defining feature is the LookML data model , a version-controlled, code-based semantic layer that defines metrics, relationships, and business logic in one place. Because the model is governed, every dashboard and report uses the same definition of revenue, active user, or churn, eliminating the spreadsheet wars where each team computes a metric differently.
Looker democratizes access in three ways:
Governed self-service — business users build their own looks, dashboards, and explores without writing SQL, while the model keeps results consistent and secure.
Dashboards and embedded analytics — dashboards are shareable, schedulable, and can be embedded into the applications customers and partners already use, so insight lives where decisions are made.
Workflows and actions — Looker can trigger alerts, send data to other systems, and feed parameterized blocks into downstream processes, turning analysis into action.
Analyzing BigQuery Data in Looker
Looker connects natively to BigQuery, which is where most Google Cloud analytics data lives. The combination creates a complete analytics pipeline.
- Stage Role Google Cloud product
- Ingest Capture events and records from apps, sensors, files Pub/Sub, Cloud Storage, Datastream
- Store and analyze Persist and query structured analytical data BigQuery
- Visualize and activate Model, dashboard, report, embed, alert Looker
Value delivered by BigQuery plus Looker:
Real-time reports — dashboards refresh from BigQuery so leaders see current numbers, not last week's export.
Consistent metrics — the LookML model ensures every dashboard computes revenue or retention the same way.
Governed access — row-level security in BigQuery and Looker's permission model ensure each user sees only what they are allowed to.
Integrated workflows — dashboards drive alerts and actions: a drop in conversion triggers a ticket, a stockout triggers a reorder.
The Value of Real-Time Streaming Analytics
Streaming analytics processes events as they arrive, so decisions are made while the event still matters. Batch analytics processes data in scheduled runs over collected windows, which is simpler and cheaper but inherently slower.
- Dimension Batch analytics Streaming analytics
- Latency Minutes to hours, often overnight Milliseconds to seconds
- Data window Fixed, completed file or table Continuous, unbounded event stream
- Cost and complexity Lower, simpler tools Higher, requires stream processing
Best for Monthly close, regulatory filings, large model training Fraud detection, IoT alerts, personalization, live ops
Business value unlocked by streaming:
Fraud detection — a payment is approved or blocked in milliseconds using a live model score, not after an overnight job.
IoT and operations — sensors on a turbine flag vibration anomalies the moment they appear, preventing failures.
Personalization — a shopping session adapts offers based on clicks within the same session.
Live operations — logistics dashboards show fleet position and ETA (estimated time of arrival) updated continuously.
Not every workload needs streaming. Use batch when latency is acceptable and cost matters; use streaming when the value of acting now is greater than the cost of the pipeline.
Modernizing Data Pipelines with Pub/Sub and Dataflow
Two Google Cloud products do most of the heavy lifting in a modern pipeline.
Pub/Sub
Pub/Sub is a managed messaging and event-ingestion service. Producers publish messages to a topic; consumers subscribe to receive them. Pub/Sub decouples producers from consumers, buffers spikes, and delivers messages at-least-once (subscribers must handle possible duplicates). It scales automatically to millions of events per second.
Role in a pipeline: a reliable, elastic front door for events from apps, devices, logs, and third-party systems.
Dataflow
Dataflow is a managed stream and batch data-processing service built on Apache Beam . It handles provisioning, scaling, and fault tolerance automatically. With Dataflow you write a pipeline once (in Java, Python, or SQL) and run it in streaming mode for live events or batch mode for backfills, using the same code.
Role in a pipeline: transformation, enrichment, windowing, and joins, the work that turns raw events into clean, structured records ready for BigQuery or downstream systems.
The End-to-End Flow
A modern Google Cloud analytics pipeline typically looks like this: Events → Pub/Sub (ingest/decouple) → Dataflow (process/enrich) → BigQuery (store/analyze) → Looker (dashboards/activate)
Pub/Sub collects events from apps, devices, and logs.
Dataflow reads from Pub/Sub, transforms and enriches the stream, and writes clean rows to BigQuery.
BigQuery stores the analytical dataset and runs the SQL.
Looker turns the BigQuery dataset into governed dashboards, alerts, and embedded analytics for the rest of the organization.
This flow is what the exam means by "smart analytics, BI tools, and streaming analytics adding value in different business use cases": each product solves one stage of the value chain, and together they move data from raw event to decision in seconds rather than overnight.
Updated-exam addition - Managed Service for Apache Spark. The exam guide effective August 12, 2026 adds a third pipeline-modernization product alongside Pub/Sub and Dataflow: Google Cloud Managed Service for Apache Spark . It runs Apache Spark workloads serverlessly, so teams that already have Spark code - typically from an on-premises Hadoop or Spark estate - can move it to Google Cloud without rewriting it in Beam and without operating a cluster. The business framing the exam wants: Dataflow is the choice when you are building a new unified stream-and-batch pipeline, while Managed Service for Apache Spark is the choice when you are modernizing an existing Spark investment and want to keep the code.