Technical Writing and Documentation Questions

Producing clear written artifacts such as design docs, runbooks, reports, specifications, and knowledge-base articles. Covers document structure and organization, writing for future readers, and keeping documentation accurate and maintainable. Distinct from live verbal communication in that it emphasizes durable, reference-quality written output.

HardTechnical

23 practiced

You must write a single-page README to explain ETL pipeline failures to both on-call engineers and product managers. Provide the structure (section headings) and the key content points for each section, ensuring the page balances technical troubleshooting steps and high-level business impact statements.

Sample Answer

EasyTechnical

30 practiced

How would you document the calculation logic for a key metric such as Monthly Active Users (MAU) to ensure other analysts and auditors can reproduce it? Include required elements: precise definition, SQL example, data sources, edge case handling, test cases, and versioning conventions.

Sample Answer

Precise definition: Monthly Active Users (MAU) = count of distinct user_id that performed at least one qualifying event within the calendar month (UTC). Qualifying events = {page_view, purchase, login}. Exclude test/internal users and bot traffic. Window: first day 00:00:00 to last day 23:59:59 UTC. Data sources: events.events_stream (event_time TIMESTAMP, user_id, event_type, device_type, is_bot BOOLEAN, env STRING) users.users_master (user_id, status, created_at) blacklist.internal_users (user_id) SQL example (Postgres-compatible): sqlWITH qualifying_events AS ( SELECT DISTINCT user_id FROM events.events_stream WHERE event_time >= '2025-06-01'::timestamptz AND event_time < '2025-07-01'::timestamptz AND event_type IN ('page_view','purchase','login') AND NOT is_bot AND env = 'prod' ), filtered_users AS ( SELECT q.user_id FROM qualifying_events q LEFT JOIN blacklist.internal_users b USING (user_id) LEFT JOIN users.users_master u USING (user_id) WHERE b.user_id IS NULL AND (u.status IS NULL OR u.status != 'suspended') ) SELECT COUNT(DISTINCT user_id) AS mau_june_2025 FROM filtered_users; Edge-case handling: Timezones: normalize all event_time to UTC before aggregations. Late-arriving events: define cut-off (e.g., +48 hours) and maintain backfill process. Duplicate IDs: dedupe using canonical user_id from users_master. Anonymous sessions: excluded unless mapped to persistent user_id. Test cases: Known test user inserted in blacklist -> must be excluded. Insert event at 2025-06-30 23:59:59 UTC -> counted. Insert event at 2025-07-01 00:00:00 UTC -> not counted. Bot flag true -> excluded. Duplicate events for same user -> count once. Versioning & governance: Store metric definition docs in repo: metrics/MAU.md with fields: name, owners, definition, SQL file path, data sources, last_modified, semantic_version (MAU v1.0.0). Use Git tags for releases. For any change, increase semver and add migration notes + impact analysis; require PR review by data owner and audit trail in CHANGELOG.

HardTechnical

32 practiced

A regulatory audit requests a clear explanation of data lineage for the metric 'monthly-active-customers' from source systems through transformations to the dashboard. Draft a written explanation suitable for auditors that traces the metric step-by-step, lists owners, notes retention windows, and calls out any known data gaps.

That is every published Technical Writing and Documentation question for Data Analyst so far. Browse the other topics in this category, or practice this one interactively.

Browse Communication, Influence & Collaboration topics Start a practice session