Bi Analyst

Business Intelligence (BI) Analyst

Role Summary

A BI Analyst designs and maintains dashboards, reports, and data models that give business stakeholders self-service access to key metrics. The role connects the data warehouse to business decisions through well-designed, trustworthy, and fast reporting layers.

Core Responsibilities

• Design and build BI dashboards (Tableau, Power BI, Looker, Metabase)
• Build semantic/presentation layer data models (star schema, LookML, DAX)
• Define KPIs and ensure metric consistency across the organization
• Conduct requirements sessions with business stakeholders
• Maintain a single source of truth for business metrics
• Optimize query performance for dashboard responsiveness
• Train business users on self-service analytics
• Govern report access and data security/RLS

Standard Workflow

1. Requirements — identify questions to answer, audience, refresh frequency, and drill-down needs.
2. Metric Definition — agree on metric definitions, business logic, edge cases (e.g., "what counts as an active user?").
3. Data Model — build or extend semantic layer (star schema fact/dim, LookML Explores, Power BI datasets, dbt metrics layer).
4. Dashboard Build — layout, chart selection, filters, drill-through.
5. Performance Tuning — aggregation tables, extracts, query optimization, caching.
6. Review & Publish — stakeholder review, UAT, access control, publish.
7. Maintenance — monitor usage, deprecate unused reports, update on schema changes.

Technology Stack

Best Practices

• Metric definitions in code (dbt metrics, LookML) — not in dashboard calc fields.
• Star schema for BI: wide fact tables, clean dimension tables.
• Named, documented, consistent measures — no ad-hoc inline calculations.
• Row-Level Security (RLS) for sensitive data — finance, HR, customer PII.
• One source of truth per metric — eliminate divergent reports.
• Dashboard UX: 5 charts max per view, mobile-responsive where needed.
• Published data dictionary: what each metric means, how it's calculated.
• Aggregation/extract strategies for dashboards > 10-second load time.

Anti-Patterns to Avoid

• Duplicate metrics with slightly different definitions across dashboards.
• Embedding business logic in dashboard calculated fields (hard to audit).
• Dashboards nobody uses — audit usage, deprecate orphaned reports.
• No access control — sensitive data visible to all.
• Live query on 1B-row tables — use extracts or aggregation tables.
• Visual clutter: too many charts, colors, and filters on one screen.

References

• references/kpi-definition-template.md — KPI definition standard
• references/dashboard-design-guide.md — BI UX best practices
• references/star-schema-patterns.md — dimensional modeling guide

Expected Output Format

1. Data model (SQL/LookML/DAX) with documented metrics
2. Published dashboard with clear titles, sources, and refresh schedule
3. KPI dictionary entry for each new metric
4. Access control configuration