Exam Targets 01–08
1. Big Data vs Data Science
Know: Big Data = management/environment challenge; Data Science = analytical/scientific practice.
Clue: scale/ingest/storage/mixed forms → Big Data; hypothesis/model/prediction/insight → Data Science.
Trap: treating them as synonyms.
Source: pp. 471–474.
2. Descriptive vs Predictive vs Prescriptive
Know: what happened/why → what likely → what action should influence outcome.
Clue: probability is predictive; recommended intervention is prescriptive.
Trap: calling every future-looking output prescriptive.
Source: pp. 474–475, 481–482.
3. Six V's as diagnostics
Know: Volume=amount; Velocity=speed; Variety=forms; Viscosity=integration/use difficulty; Volatility=change/useful life; Veracity=trust.
Trap: knowing the words but missing the management implication.
Source: pp. 476–477.
4. Nine activities vs seven-step loop
Know: broader context diagram has nine management activities; seven-step figure focuses model-development loop. Environment establishment and communication are explicit in the broader view.
Trap: forcing the diagrams into one identical list.
Source: pp. 473, 475–476.
5. ETL vs ELT
Know: transform/integrate before load vs land first and transform/integrate for use later.
Trap: ELT ≠ no transformation or governance.
Source: pp. 471–472, 476–478.
6. Data lake vs data swamp
Know: raw/flexible data can be managed. Swamp = missing context, inventory, provenance, quality, relationships, ownership.
Trap: raw = swamp.
Source: pp. 478–480, 499–500.
7. Batch vs Speed vs Serving
Know: batch=history/completeness; speed=current/low latency; serving=merged/exposed view.
Clue: business timing and completeness requirement.
Trap: fastest is always best.
Source: pp. 479–480.
8. Supervised vs Unsupervised vs Reinforcement
Know: labeled target vs hidden-pattern discovery vs goal/reward feedback.
Trap: naming clusters afterward does not make original learning supervised.
Source: pp. 480–481.