CompTIA Data+ DA0-002 · Free study guide

Domain 3 — Data Analysis (24%)

Data Analysis is the largest DA0-002 V2 domain. Its current scope has three parts: choose a communication approach, select statistical methods and common functions, and troubleshoot analysis issues. This supplemental review connects those parts into one workflow. The objective lessons remain the complete checklist.

Start with the decision and audience

An analysis can be mathematically correct and still fail if it answers the wrong question, exposes unnecessary detail, or cannot be understood by its audience. Before building, confirm the decision, metric definition, population, time window, comparison, required detail, refresh need, and access boundary.

A mock-up makes those requirements concrete before implementation. It can show proposed KPIs, filters, labels, visual positions, and navigation without a live connection. When two stakeholders interpret “active customer” differently, the mock-up exposes the disagreement while it is still inexpensive to fix.

Match the result to the persona:

• A C-suite audience generally needs a few strategic KPIs, target gaps, risks, implications, and a clear decision request.
• An individual contributor needs actionable records, filters, definitions, and enough detail to complete the work.
• A technical reviewer needs sources, logic, assumptions, limitations, and diagnostics that make the analysis reproducible.
• A non-technical audience needs plain language, familiar units, concise context, and the consequence of the result.
• An external audience receives only approved scope and detail, with disclosure and access controls applied before delivery.

Internal does not mean unrestricted. Minimize sensitive information and enforce row- or column-level access at the data path rather than merely hiding a field in a visual. Support visual accessibility with readable type, contrast, labels, logical reading order, keyboard access, and a non-color cue for every critical status. Support auditory accessibility with captions, transcripts, and visible counterparts for sound-only alerts.

Classify the analytical approach

DA0-002 V2 names four approaches:

Classify the work, not the product label. A dashboard that displays last month’s values is descriptive. A model estimating next month’s demand is predictive. A system recommending inventory quantities under cost and service constraints is prescriptive. An analysis using a representative sample to estimate a population value is inferential.

Choose center and spread for the data

The mean is the sum divided by the observation count and uses every numeric value. The median is the middle value after sorting and is less affected by extreme observations. The mode is the most frequent value and can describe numeric or categorical data. A data set may have one mode, several modes, or no mode under a chosen convention.

For 10, 12, 12, 13, 15, and 78 minute response times, the mean is about 23.3, the median is 12.5, and the mode is 12. The high value pulls the mean away from most observations, so the median better represents the center of this skewed sample. That does not make 78 invalid; it is evidence to investigate.

The range is maximum minus minimum. Variance summarizes squared distance from the mean, while standard deviation is its square root and returns spread to the original unit. Equal means do not establish equal distributions: two groups can share a mean while having very different spread, skew, or subgroups.

Percentiles locate values within an ordered distribution. Quartiles mark the 25th, 50th, and 75th percentiles, and the interquartile range is Q3 - Q1. A common screening rule flags values below Q1 - 1.5 × IQR or above Q3 + 1.5 × IQR. A flag identifies a candidate for review, not automatic permission to delete it.

Apply logical, mathematical, date, and string functions

Functions translate business rules into repeatable operations. IF or CASE classifies records. AND requires all joined conditions, OR requires at least one, and NOT reverses a condition. Parentheses make mixed conditions explicit. Nulls need null-aware logic; ordinary equality does not treat an unknown value like a known value.

Mathematical functions such as ABS, ROUND, FLOOR, CEILING, POWER, and SQRT operate on values. Aggregates such as SUM, COUNT, MIN, MAX, and AVG operate across rows. Round at the reporting boundary when possible, because repeated intermediate rounding can accumulate error.

Date operations add or subtract intervals, calculate durations, extract parts, or group timestamps into periods. Resolve time zones before comparing instants, and distinguish a calendar date from a timestamp. String functions such as TRIM, UPPER, LOWER, LENGTH, SUBSTRING, REPLACE, and concatenation standardize or reshape text. They do not establish identity: two normalized names may still describe different people.

Troubleshoot with evidence

Begin by recording expected and observed values, user, filters, period, time, and environment. Reproduce the symptom, identify recent changes, and isolate connectivity, source, transformation, query, and presentation layers. Change one variable at a time, apply the smallest defensible repair, then reconcile the result and monitor the next run.

Connectivity, authentication, and authorization are distinct. A timeout can indicate network path or service availability. Invalid credentials point to identity or token problems. Permission denied can mean authentication succeeded but the identity lacks access.

A SQL syntax error prevents execution; a logic error runs and returns a wrong result. Check join cardinality, row counts, key uniqueness, aggregation grain, null handling, filter placement, and mixed AND/OR conditions. For corrupt or unexpectedly shaped input, preserve and quarantine the original, then compare schema, encoding, delimiter, column count, required fields, row count, and control totals with a known-good run.

Logs should identify run, time, environment, version, stage counts, duration, and sanitized errors without exposing secrets. Vendor documentation and release notes establish version-specific behavior; community suggestions are hypotheses to verify safely, not authority by themselves.

What to drill

For 3.1, justify a communication choice from audience, decision, accessibility, sensitivity, KPI, and detail requirements. For 3.2, classify the four current analysis approaches, choose measures that fit a distribution, interpret spread and outliers, and apply logical, mathematical, date, and string operations. For 3.3, practice tracing a discrepancy from symptom to source and proving the repair with known results.