Rapid Recall Key 01–20

1. Big Data: large/fast/varied and difficult-to-manage data whose characteristics require different storage, integration, Metadata, DQ, and analytical approaches; not a fixed row/byte threshold.
2. Data Science: iterative scientific/analytical practice using mining, statistics, ML, integration, and modeling to discover patterns, test hypotheses, and build useful predictive/prescriptive insight.
3. “Big” is relative: management difficulty depends on organization, technology, data type, and workload; no universal size threshold.
4. Six V's: Volume, Velocity, Variety/Variability, Viscosity, Volatility, Veracity.
5. Volume vs Velocity: amount/scale vs speed of generation/capture/share/use.
6. Variety vs Viscosity: different forms/structures vs difficulty using/integrating the data.
7. Volatility vs Veracity: change/useful life vs trustworthiness.
8. Four goals: discover data-business relationships; discover/analyze new factors; package model outputs for stakeholders; integrate organizational practices with DM/Big Data/Data Science best practices.
9. Nine activities: strategy/business needs; environment; choose sources; acquire/ingest; hypotheses/methods; integrate/align; explore/model; communicate; deploy/monitor.
10. Seven phases: strategy/business need; choose sources; acquire/ingest; hypotheses/methods; integrate/align; explore/model; deploy/monitor.
11. Why compatible: the context diagram is broader and separately exposes environment establishment and communication; the seven-step view focuses model development.
12. Descriptive: what happened and why?
13. Predictive: what is likely to happen?
14. Prescriptive: what action should be taken to influence the outcome?
15. Rear-view vs windshield: traditional BI looks backward; Data Science adds forward-looking predictive/prescriptive insight.
16. ETL vs ELT: transform/integrate before load vs load/ingest first and transform/integrate later.
17. Why ELT: diverse raw sources can be landed before a specific analytical use determines needed transformation/integration.
18. Data lake: flexible environment for varied data used for exploration, detailed history, archive, streams, or analytics.
19. Swamp: missing inventory, Metadata, provenance, DQ, relationships, ownership, and discoverability make the lake opaque/untrusted.
20. At ingest: preserve origin/source, size, currency, content/structure, lineage, and other Metadata plus profiling/DQ context.

← Prompts 01–20 · Key 21–40 →