DP-900 Microsoft Azure Data Fundamentals Roadmap

Start with what the data is and what work must happen, not with a product name.

• Distinguish structured tables, semi-structured JSON or XML, and unstructured text, images, audio, and video.
• Compare CSV, JSON, and columnar analytical formats such as Parquet from schema, nesting, readability, compression, and selective-read needs.
• Separate files, object stores, file shares, databases, data lakes, and warehouses by access pattern.
• Contrast transactional point operations with analytical scans, grouping, aggregation, and historical trends.
• Map database administrator, data engineer, and data analyst responsibilities without assuming the roles never overlap.

Learn the relational model before comparing managed engines.

• Identify entities, tables, rows, columns, primary keys, foreign keys, constraints, indexes, and views.
• Normalize repeated entities to reduce redundancy and update anomalies; keep the intended transaction and reporting needs visible.
• Recognize SELECT, INSERT, UPDATE, DELETE, CREATE, and common filtering and join purposes.
• Choose Azure SQL Database for managed database-scoped needs, Managed Instance for broad instance compatibility, and SQL Server on Azure VMs when host and full instance control are required.
• Identify Azure Database for PostgreSQL and Azure Database for MySQL when application engine compatibility calls for them.
• Explain that more control usually means more customer administration.

Choose non-relational stores from object, file, key-attribute, document, graph, and distribution requirements.

• Use Blob Storage for object data and distinguish block, append, and page-blob patterns.
• Use Azure Files when applications or users need managed SMB or supported NFS file shares.
• Use Table storage for schemaless key-attribute entities with access patterns centered on partition and row keys.
• Use Azure Cosmos DB for globally distributed operational data where latency, scale, consistency, partitioning, and API fit justify it.
• Differentiate Cosmos DB for NoSQL, MongoDB, Apache Cassandra, Apache Gremlin, and Table compatibility from the application's model and protocol.
• Document retention, authorization, encryption, network access, backup, and cost for every choice.

Trace data from source through ingestion, processing, analytical storage, and serving.

• Define batch boundaries, incremental ingestion, schema validation, transformation, reconciliation, lineage, and quality handling.
• Compare data lakes for broad raw and curated data with warehouses for structured analytical serving.
• Describe Azure Databricks as a collaborative Spark and lakehouse analytics platform.
• Describe Microsoft Fabric as a unified SaaS analytics platform spanning integration, engineering, warehousing, real-time intelligence, data science, and Power BI.
• Differentiate batch processing from low-latency event processing.
• Recognize Azure Stream Analytics and Fabric eventstreams or Real-Time Intelligence as Microsoft options for real-time scenarios.

Connect analytical models to responsible visual communication.

• Use Power BI to connect, transform, model, visualize, and distribute governed insight.
• Build a star model with a declared fact grain, descriptive dimensions, relationships, and reusable measures.
• Choose line charts for ordered trends, bars for discrete comparisons, scatter plots for relationships, cards for bounded headlines, and tables for exact detail.
• Use slicers for interactive filtering and validate blank, multi-select, and security behavior.
• Complete all three projects, review 40 cards, and answer the exact 14/12/9/15 bank allocation.
• Take Microsoft's official Practice Assessment and use the exam sandbox; check the live study guide again before booking.