Menu
Data Analyst
Three components - Automation Resistance, Structural Moat, and Demand - add up to 39.
Federal labor data does not isolate this job; the workforce and openings numbers here come from a broader data-science-and-analytics occupation. That row includes modeling-heavy data scientist work, so the numbers are a directional public comparison rather than an exact data-analyst count.
Routine query, dashboard, and summary work is exposed, while business-question judgment keeps a narrower human lane for quality checks, metric definitions, stakeholder context, data-quality review, ambiguity, source skepticism, and responsible translation into decisions under pressure.
Observed AI exposure is about 46.1%, and modeled median job-loss pressure is about 37.2% in the broader data-science-and-analytics occupation. That matches the role: generated Structured Query Language (SQL), dashboards, summaries, and anomaly explanations reach the routine layer directly. Business context and metric judgment keep some work human.
AI is useful for draft queries, chart suggestions, table summaries, and data checks. The gain often lets teams produce reporting faster with the same or fewer people, so the worker benefit is partial. Analysts gain more when they use AI to test questions and verify data, not just to generate dashboards.
The formal moat is thin because the work is screen-based and unlicensed; practical protection comes from trusted metrics, domain context, access to messy internal data, and being close enough to the business to know what a number means.
The work is screen-based analytics. There is no physical setting, field condition, or hands-on environment that protects the seat. Any durability has to come from business context, trusted data access, and judgment about what a metric actually means.
There is no broad license for data analysts. Privacy, governance, and compliance rules can shape the work, but they do not create a legal entry gate. Employers may prefer degrees, certificates, or tool experience, yet those are hiring signals rather than enforceable protection.
Physical robotics does not replace this role because the work is data interpretation and decision support. The real pressure is software automation inside spreadsheets, databases, and business-intelligence tools, which is already counted in the automation component.
The broader federal occupation sits in a higher-preparation zone, usually tied to a bachelor-level path. Data analysts can enter through degrees, certificates, or portfolios, but better roles usually require enough statistics, SQL, business context, and communication skill to handle ambiguous questions.
Demand benefits from a growing broader data row, but that row is wider than business-intelligence reporting, and routine analyst work is compressible when dashboards, queries, summaries, and variance explanations become easier to generate quickly at scale.
Public tables do not isolate this job; the broader data-science-and-analytics occupation has about 245,900 jobs and about 23,400 annual openings. That gives strong public scale, but it is not an exact count of business-intelligence analyst seats.
The source fit is mixed because the public row includes modeling-heavy data scientist work. It is useful for scale and wages, but the business-intelligence and reporting lane has a different task mix: dashboards, metric definitions, stakeholder questions, and recurring business reports.
Resilience is limited by direct AI compression of SQL, dashboards, summaries, and first-pass explanations. The work is more resilient when analysts own metric definitions, data quality, and decision translation. It is weaker when the job is mostly producing routine reporting artifacts.
The case weakens if business users can reliably ask questions, generate dashboards, and receive usable summaries without an analyst. The exposed roles would be scheduled reporting and simple variance explanations with little metric ownership, data-quality accountability, or decision context inside teams.
The case strengthens if companies find that AI-generated answers create bad definitions, misleading dashboards, or data-quality errors. Analysts who can audit metrics, challenge assumptions, explain uncertainty, and translate findings into decisions would become more valuable inside teams with real authority.
A mixed outcome needs review if embedded business analysts stay useful while central report-production roles shrink. Readers would need to compare roles by stakeholder access, data ownership, recurring business context, and decision responsibility rather than title alone during hiring decisions.