CompTIA Data+ DA0-002 (V2) Exam Practice Test 3

KPIs are the specific, pre-agreed metrics identified as the most important indicators of business health, such as monthly recurring revenue and churn rate. A mock-up is an early draft prototype, user persona type concerns role-based content tailoring, accessibility concerns usability for people with disabilities, and level of detail concerns content granularity rather than describing these specific named business metrics.

A stress test evaluates system performance and stability under heavy simulated load, exactly matching this thousands-of-simultaneous-users scenario, distinct from a unit test which checks individual code components in isolation.

Encrypting data while it moves between systems is encryption in transit, while encrypting it while stored on disk is encryption at rest. Masking obscures values reversibly rather than encrypting them, anonymization permanently removes identifying information, and role-based access control restricts who can view data rather than encrypting it.

Metadata is descriptive information about data itself, such as creator, last update time, and format, without containing the actual data values. Source of truth identifies the authoritative dataset among possibly conflicting copies, data versioning tracks changes over time, data lineage traces origin and transformation history, and a data dictionary defines field-level meaning specifically rather than this broader descriptive dataset-level information.

Monitoring alerts specifically for data structure changes would proactively flag that the source table's structure was modified, before the report's numbers became inaccurate. Data filtering addresses which rows are shown, peer review of calc logic checks calculation formulas, source validation alone confirms the source is online and accessible without necessarily flagging structural changes, and code review alone examines report code rather than monitoring the underlying source schema for changes.

Slow loading caused specifically by querying millions of raw rows rather than a pre-aggregated summary is a large data size issue. Filter not working correctly concerns a filtering mechanism malfunction, stale data concerns outdated figures, corrupt data concerns damaged values, and slow refresh rate concerns the speed of an active data refresh rather than the underlying data volume driving load time.

A dynamic dashboard automatically updates its underlying data as new information arrives, showing current figures at any moment. A static dashboard shows a fixed view that does not automatically update, an executive summary is a condensed document rather than a live dashboard, a self-service portal describes user-driven data access broadly, and ad hoc describes on-demand request timing rather than automatic live updating.

A legend explains what different colors or symbols represent across multiple data series within a visual, exactly matching this region-color key. Labels identify specific data points or axis values, branding applies company visual identity, color schemes generically concern color selection and consistency broadly, and infographics are a visual type rather than this symbol-explanation design element.

A chart plots numeric data to show trends or comparisons, a map visualizes geographic data, a pivot table allows interactive drag-and-drop summarization, and an infographic combines multiple visual elements and text into one polished, easily digestible graphic, each serving a distinct visualization purpose.

Validating the data source confirms the original system is online, accessible, and providing data in the expected format before investigating further downstream. Enabling logging captures execution detail during processing, consulting vendor communities seeks external expertise, and basic SQL code review and corrupted data checks focus on query syntax or data values rather than confirming the upstream source itself is functioning.

A typo in a column name causing a query to fail is a basic SQL code issue. Connectivity-related issues concern failed connections to the database itself, user-reported issues originate from an end user flagging a problem regardless of cause, and corrupted data and stale data concern the state of the data itself rather than a syntax error in the query.

Inferential statistics draw conclusions about a larger population based on a smaller sample, accounting for margin of error and confidence, exactly matching this scenario. Descriptive statistics summarize the sample itself without generalizing to the population, predictive statistics forecast future outcomes, prescriptive methods recommend specific actions, and diagnostic is not one of the four basic statistical method categories in this objective.

Prescriptive methods recommend a specific action to take based on data and modeling, exactly matching recommending the best next action for a specific customer. Descriptive statistics summarize past data, predictive statistics forecast future outcomes without necessarily recommending a specific action, inferential methods draw conclusions about a population from a sample, and diagnostic is not one of the four basic statistical method categories in this objective.

varchar stores variable-length text without padding shorter values to a fixed length, unlike char which reserves fixed-length storage. nvarchar specifically adds Unicode character support on top of the variable-length behavior, numeric stores numbers, and Boolean stores true/false values rather than variable-length text.

Ensuring content is appropriate for an audience outside the company, distinct from internal-only content, is the internal versus external consideration. Sensitive versus non-sensitive concerns confidentiality classification more broadly regardless of internal or external audience, technical versus non-technical audience concerns language complexity, level of detail concerns granularity, and accessibility concerns usability for people with disabilities.

Augmentation adds new external data to enrich an existing dataset with fields not originally collected. Merging generically combines datasets based on a common key without specifically implying external enrichment, appending generically adds new rows to an existing structure, derived variables create new calculated fields from existing data, and deletion removes records or fields rather than adding new external data.

Conversion changes a value's format or data type to a consistent standard, exactly matching standardizing inconsistent date formats. Binning groups continuous values into categories, clustering groups similar records together, exploding expands array fields into multiple rows, and augmentation adds new data to enrich a dataset rather than standardizing existing format inconsistencies.

The same value unnecessarily repeated across multiple unrelated systems rather than maintained in one place is redundancy, while a logically impossible value like a negative price is an outlier. Duplication describes the same entity appearing as separate records, missing values describes absent required data, and completeness describes whether all expected data is present rather than either of these two described patterns.

Completeness confirms that all expected fields are populated for every record, with no gaps in required information. Validation checks conformance to formats, ranges, and business rules more broadly, duplication describes repeated entity records, redundancy describes unnecessarily repeated data, and outliers describes unusual extreme values rather than checking for populated required fields.

ELT loads raw data directly into the target system first, then performs transformation afterward using that system's processing power. ETL instead transforms data before loading it into the target system, nested queries embed one query inside another, data integration generically combines sources without specifying this load-then-transform order, and temporary tables store intermediate results rather than describing this pipeline pattern.

Data integration combines data from multiple different source systems into a single unified dataset, reconciling differences between them. Querying generically retrieves data using structured requests, basic query optimization improves query performance, ETL specifically describes a broader structured extract-transform-load pipeline process, and data collection describes gathering new data directly rather than combining existing separate sources.

A foundational model is trained on vast general data and can be adapted to many specific downstream tasks without retraining from scratch. Robotic process automation automates rule-based tasks, natural language processing generically describes working with language broadly, deep learning generically describes the underlying neural network technique, and automated reporting is a specific RPA use case rather than describing this adaptable general-purpose base model.

Notebooks combine executable code cells, visualizations, and formatted text in a single shareable document, ideal for exploratory analysis, distinct from a plain text editor or a general-purpose IDE.

Object storage stores files with associated metadata, accessed via API, optimized for massive scalability rather than fast random access to small file pieces. Block storage provides raw, low-latency volumes for fine-grained access, file storage provides a hierarchical directory structure, and local and shared storage describe physical or network attachment rather than this metadata-rich, API-accessed storage model.

APIs provide structured, defined endpoints that return data in a consistent format directly from a running application. Files are static exports rather than live requests, logs record system or application events, web scraping extracts data from rendered web pages rather than structured endpoints, and data repositories describe broader storage systems rather than this specific request-based access method.