CompTIA Data+ DA0-002 · Free study guide

Domain 4 — Visualization and Reporting (20%)

Domain 4 connects visual choice, report delivery, and report validation. The goal is not merely to recognize chart names: it is to select an honest visual, deliver it through a channel that fits the audience and cadence, and trace a broken report back to evidence.

Chart selection: match the question, not the data

Start with the question the visual needs to answer:

• Change over time → line chart. Trend, slope, inflection — 36 months of monthly active users is a line, never a pie.
• Compare categories → bar chart. Human perception judges bar length with high precision, which is also why the bar family carries strict honesty rules (below).
• Distribution of one continuous variable → histogram. The bar/histogram distinction is a core design choice: bar charts compare discrete categories (spaced, reorderable bars); histograms bin one continuous variable (touching bars, fixed order).
• Relationship between two numeric variables → scatter plot. Square footage versus sale price across thousands of homes. Add a third numeric variable by sizing the points and it becomes a bubble chart.
• Compare distributions across groups → box plots side by side: median, IQR, whiskers, and outliers for five server clusters in one visual.
• Intensity across two categorical axes → heat map. Weekday × hour call volume is a useful scenario.
• Part-to-whole → pie chart, only with a handful of categories. Past five or six slices, angles become hard to compare; a sorted bar chart is usually clearer. 3-D pies distort apparent area and should be avoided.
• Totals plus composition → stacked bar: each region's total bar divided into product-line segments.
• Bridge from a start value to an end value through gains and losses → waterfall chart. Last year's profit to this year's through pricing, volume, cost, and currency effects — a natural use for this chart.

Honest visuals

Three rules help keep visual communication honest:

Bar axes generally start at zero. Bars encode value as length; truncating the baseline exaggerates small differences — a 45% versus 47% market share drawn from a 44% baseline looks like a threefold gap. Line charts can sometimes justify a non-zero baseline. A specialized bar chart that breaks an axis must make the break unmistakable and still support an honest comparison.

Dual axes invite manipulation. Two series on independently scaled y-axes can be tuned to cross anywhere and imply any relationship. The risk is spurious correlation manufactured by axis choice; the safer designs are indexed scales or separate aligned panels.

Annotate the anomaly. A revenue dip caused by a warehouse outage should carry an annotation at that point, so future readers do not misread it as demand collapse. Smoothing the dip away misrepresents the data; explaining it in place is data storytelling done right.

Accessibility

One rule dominates: never encode meaning in color alone, especially with red and green, because red-green deficiencies are common. The fix is a second channel — direct labels, icons, patterns, or a colorblind-safe palette. Related good practice: sufficient contrast, readable font sizes, and sequential palettes for ordered data. Color remains a legitimate encoding; it just cannot be the only one.

Dashboards

Dashboard design depends on layout and interactivity conventions:

• Headline KPIs go at the top. Viewers scan from the top; the summary numbers the dashboard exists to deliver belong where the eye lands first, with supporting detail below.
• Drill-down navigates a hierarchy from summary to detail: nation → region → state → account. Recognize the term from the interaction described.
• Self-service filtering — parameters, slicers, date-range and territory filters — lets consumers reshape a governed view themselves instead of filing a report request for each new question.

Reports: form follows audience and cadence

Static versus dynamic. A static report is a frozen point-in-time snapshot; a dynamic dashboard can update on demand, on a schedule, or from a real-time feed. Figures submitted for a closed regulatory period may need an immutable, versioned snapshot rather than a link whose values can change after submission.

Audience calibration. Executives get headline findings and KPI movement first, detail in an appendix. Raw row dumps, SQL listings, and maximal jargon all fail the audience — rigor is demonstrated by clarity, not density.

Delivery cadence. Recurring operational needs get scheduled reports — inventory figures generated and distributed automatically at 7:00 a.m. daily. Ad hoc reports answer one-off questions. Matching cadence to need is the whole question.

Validate and troubleshoot the report

Start by separating symptoms. Load time is how long the report takes to open or respond. Refresh rate is how often a refresh is scheduled. Freshness latency is the delay between a source event and its appearance in the report. A large data set may require aggregation, narrower queries, or a different delivery method; increasing refresh frequency does not make a slow query efficient.

Reproduce broken filters with known inputs and compare filtered totals with a source control. For stale or corrupt results, inspect source timestamps, row counts, schemas, pipeline logs, encodings, and refresh history before changing the visible calculation. Review query and calculation code, have a peer check high-impact changes, and reconcile output to an independent trusted total. Schema-contract checks and monitoring alerts make failures visible before a consumer discovers them. After a repair, rerun the original case and adjacent filters so the fix does not introduce a new defect.

What to drill

Practice the full decision chain: choose a visual for the question; add titles, units, labels, legends, annotations, branding, and accessible encodings; choose a static, dynamic, recurring, ad hoc, snapshot, real-time, summary, dashboard, or portal delivery method; then validate filters, calculations, source data, freshness, performance, and monitoring evidence.