Big Data: A Very Short Introduction (Very Short Introductions)
One book, placed in its field — what this book gets right, where it goes further than the rest, and what the rest of the field adds.
Mike West, Editor · Leading With Data · peopleanalyst.com/guides/profile-big_data_a_very_short_introduction
Arrow keys or swipe to move · N for notes · F for full screen
Big Data: A Very Short Introduction (Very Short Introductions)
One book, placed in its field — what this book gets right, where it goes further than the rest, and what the rest of the field adds.
Mike West, Editor · Leading With Data · peopleanalyst.com/guides/profile-big_data_a_very_short_introduction
The problem
Privacy and security risk erodes when it is left to instinct instead of method.
You were taught the moves piecemeal, never the whole model.
Big Data: A Very Short Introduction (Very Short Introductions) · Mike West, Editor · Leading With Data
The model
- 01Data Volume, Velocity, and Variety
- 02Data Veracity (Quality)
- 03Distributed Storage Architecture
- 04Analytic Technique Application
- 05Useful Information Extraction
- 06Predictive Model Accuracy
- 07Data Security and Privacy Measures
- 08Decision Quality and Organizational Value
Big Data: A Very Short Introduction (Very Short Introductions) · Mike West, Editor · Leading With Data
Step 1 of 5 · Data Volume, Velocity, and Variety
Big data is defined by the simultaneous pressure of scale, speed, and heterogeneity — any single dimension alone rarely requires specialized systems.
Misconception
Practitioners assume big data is fundamentally about volume — that once you have enough terabytes, you have a big data problem.
Reality
A modest-volume feed of high-velocity, heterogeneous sensor and social data is often harder to handle than a large static archive; variety and velocity break traditional relational tools long before raw size does.
Big Data: A Very Short Introduction (Very Short Introductions) · Mike West, Editor · Leading With Data
Step 2 of 5 · Data Veracity (Quality)
More data does not fix bias — it entrenches it, so provenance matters more as volume grows.
Misconception
Practitioners believe that huge sample sizes wash out noise and bias, so cleaning becomes unnecessary at big-data scale.
Reality
Volume amplifies systematic bias rather than diluting it; a skewed collection process produces more of the same skew, and more data only makes a wrong signal look statistically stronger.
Big Data: A Very Short Introduction (Very Short Introductions) · Mike West, Editor · Leading With Data
Step 3 of 5 · Distributed Storage Architecture
Horizontal scalability is bought with relaxed consistency, not for free.
Misconception
Practitioners think switching to a distributed or NoSQL system is mainly a performance upgrade that keeps the familiar querying and consistency of a relational database.
Reality
Distributed and NoSQL systems trade the strict consistency and rich joins of relational databases for scalability and fault tolerance; you gain horizontal growth by accepting eventual consistency and schema flexibility, which changes how you must design queries.
Big Data: A Very Short Introduction (Very Short Introductions) · Mike West, Editor · Leading With Data
Step 4 of 5 · Analytic Technique Application
The analytic question determines the technique — technique should never precede the question.
Misconception
Practitioners treat machine learning as a single interchangeable tool and expect that throwing any algorithm at data will surface whatever is useful.
Reality
Each technique encodes a specific assumption about structure — clustering finds groups, classification assigns labels, PageRank ranks by linkage — and choosing one commits you to its worldview; the algorithm cannot find patterns its formulation is blind to.
Big Data: A Very Short Introduction (Very Short Introductions) · Mike West, Editor · Leading With Data
Step 5 of 5 · Useful Information Extraction
Extraction is the hinge between all your inputs and every downstream benefit — it fails silently when any upstream element is weak.
Misconception
Practitioners assume that once data is collected and stored, the useful information is already 'in there' waiting to be queried out.
Reality
Information is manufactured, not retrieved — it emerges only through the combination of appropriate storage, deliberate technique, and clean input, and poor choices anywhere upstream mean the pattern simply never forms.
Big Data: A Very Short Introduction (Very Short Introductions) · Mike West, Editor · Leading With Data
A framework to use
The Four 'V's of Big Data Characterization
This framework is part of the full guide. See the guide →
Big Data: A Very Short Introduction (Very Short Introductions) · Mike West, Editor · Leading With Data
The cheat sheet
- Data Volume, Velocity, and VarietyBig data is defined by the simultaneous pressure of scale, speed, and heterogeneity — any single dimension alone rarely requires specialized systems.
- Data Veracity (Quality)More data does not fix bias — it entrenches it, so provenance matters more as volume grows.
- Distributed Storage ArchitectureHorizontal scalability is bought with relaxed consistency, not for free.
- Analytic Technique ApplicationThe analytic question determines the technique — technique should never precede the question.
- Useful Information ExtractionExtraction is the hinge between all your inputs and every downstream benefit — it fails silently when any upstream element is weak.
- Predictive Model AccuracyFit to past data is not proof of future accuracy — validate on data the model has not seen.
- Data Security and Privacy MeasuresSecurity is layered defense; no single control is sufficient at big-data scale.
- Decision Quality and Organizational ValueValue is realized at the point of decision, not at the point of insight.
Big Data: A Very Short Introduction (Very Short Introductions) · Mike West, Editor · Leading With Data
Sources
- Big Data_ A Very Short Introduction (Very Short Introductions)
Big Data: A Very Short Introduction (Very Short Introductions) · Mike West, Editor · Leading With Data