Learn how modern data platforms are structured. This subtheme covers the architecture of data systems, data pipelines, and the patterns used to organize and process information at scale.
Flashcards combined with spaced repetition improve active recall. You review at the right time, retain knowledge longer, and track progress card by card.
Card 1
Consistent reuse of trusted data across teams and use cases.
Explanation
Seeing data as a shared asset drives common standards, shared infrastructure, and cross-domain reuse.
Common mistake
Thinking data belongs only to the application that first created it, instead of the whole organization.
Card 2
Storage and processing can scale independently.
Explanation
Independent scaling helps control cost and performance by adjusting only the constrained resource.
Common mistake
Assuming separation of storage and compute automatically removes all performance bottlenecks.
Card 3
Data must fit a predefined structure before storage.
Explanation
Schema-on-write enforces validation at ingestion, which simplifies downstream querying but reduces flexibility.
Common mistake
Confusing schema-on-write with having no schema and total flexibility at storage time.
Card 4
Serving responses even if data is temporarily stale.
Explanation
Favoring availability accepts eventual, not immediate, consistency when partitions or delays occur.
Common mistake
Believing high availability always comes with perfectly up-to-date, strongly consistent data.
Card 5
It requires embedding controls for access, quality, and compliance directly into data systems.
Explanation
Architecture must support governance with capabilities like authorization, validation, and auditing.
Common mistake
Treating governance as a purely bureaucratic activity disconnected from platform design.
Card 6
It shows how data was produced, transformed, and used.
Explanation
Lineage makes it possible to trace data from source to consumption, which supports debugging, impact analysis, compliance, trust, and controlled change management.
Common mistake
Assuming lineage can be reconstructed later without designing systems and metadata to capture it properly.
Card 7
Helping users find and understand datasets through metadata.
Explanation
A data catalog improves discoverability by centralizing metadata such as definitions, owners, tags, and usage context, so users can find and understand available datasets.
Common mistake
Treating a catalog like a storage tool rather than a discovery and metadata layer.
Card 8
Operational systems run transactions; analytical systems support reporting and analysis.
Explanation
Transactional workloads optimize for many small updates; analytical workloads optimize for large read-heavy queries.
Common mistake
Designing a single database to efficiently handle both heavy analytics and critical transactions.
Card 9
To store newly ingested data with minimal processing.
Explanation
A landing zone is the initial area where incoming data is captured with minimal alteration, preserving source fidelity before validation, transformation, and downstream modeling.
Common mistake
Skipping the landing zone and loading unvalidated source data directly into curated consumption layers.
Card 10
Cleaned, modeled, business-ready data for broad use.
Explanation
The curated layer contains cleaned, structured, and business-aligned data designed for reliable reuse across reporting, analytics, and self-service use cases.
Common mistake
Confusing curated data with raw ingestion outputs or assuming curation means copying source tables as-is.
Create your Memia account to unlock this deck and start focused practice sessions with progress tracking.