Data Workflow Analyst Skill
Clean, validate, transform, analyze, and summarize spreadsheet, CSV, JSON, and database-style data with clear workflows, quality checks, formulas, insights, and report-ready outputs.
Data Workflow Analyst Skill is a free, reviewed AI skill for data workflows. Clean, validate, transform, analyze, and summarize spreadsheet, CSV, JSON, and database-style data with clear workflows, quality checks, formulas, insights, and report-ready outputs. It works with ChatGPT, Claude, Gemini and is ready to use out of the box.
- • The skill cannot analyze data that has not been provided or described with enough detail.
- • It should not invent rows, columns, values, calculations, trends, correlations, or conclusions.
- • Data quality issues may limit the reliability of summaries and insights.
About this skill
The Data Workflow Analyst Skill helps users work with messy data from spreadsheets, CSV files, exports, forms, dashboards, databases, and APIs. It guides cleaning of inconsistent data, detection of missing values, standardization of columns, validation of records, and creation of clear summaries and analysis reports. The skill is useful for students, business users, analysts, founders, operations teams, marketers, researchers, and AI coding users who need structured data work without losing accuracy or traceability. It never invents data, calculations, or trends, always separates confirmed findings from assumptions, and consistently recommends backups, validation checks, and expert review for high-stakes decisions. Whether the task is a quick spreadsheet cleanup or planning a repeatable Python or SQL workflow, the skill keeps the process careful, transparent, and privacy-aware.
What it does
This skill helps users clean and validate spreadsheet, CSV, JSON, and database-style data by inspecting columns and schema, detecting missing values and duplicates, defining cleaning rules with before-and-after examples, and validating transformations against row counts and expected totals. It also helps create descriptive summaries, trend and comparison analyses, data quality reports, and dashboard preparation plans, while explaining spreadsheet formulas, drafting SQL query plans, and outlining Python or pandas workflows, always separating observed findings from assumptions and flagging data quality limitations before conclusions are drawn.
What is included
- SKILL.md — concise runtime instructions for the AI assistant
- workflow.md — step-by-step workflow for cleaning, validating, analyzing, and reporting on data
- data-cleaning-framework.md — framework for data inventory, quality issues, cleaning rules, transformations, validation, and documentation
- data-validation-checklist.md — checklist for structure, missing data, duplicates, formats, ranges, calculations, privacy, and readiness scoring
- analysis-and-reporting-guide.md — guide for summaries, trends, comparisons, quality reports, dashboards, insight writing, and report structure
- output-templates.md — reusable formats for cleaning plans, quality reports, analysis summaries, formulas, SQL plans, pandas workflows, and dashboard preparation
- examples.md — realistic input and output examples for data cleaning, validation, analysis, and reporting
How to use it
1. Download the ZIP file for this skill 2. Extract the files to a folder on your computer 3. Open your AI assistant, spreadsheet assistant, coding assistant, or data analysis tool 4. Upload or paste the skill files if the tool supports it 5. Share the dataset, sample rows, schema, column names, or analysis goal 6. Redact private or sensitive data before sharing anything 7. Ask the assistant to apply the Data Workflow Analyst Skill
Examples
I have a CSV export of customer requests. Columns: request_id, customer_name, email, category, status, created_at, completed_at. The status values are inconsistent: done, Done, completed, in progress, In Progress, received. I want to clean the data and summarize how many requests are received, in progress, and done.
Data cleaning plan: Standardize the status column into three values: Received, In Progress, and Done. Map "done", "Done", and "completed" to Done. Map "in progress" and "In Progress" to In Progress. Map "received" to Received. Validation checks: Confirm every row has a request_id. Count missing emails separately instead of deleting them. Check whether completed_at is filled only for Done requests. Compare row count before and after cleaning to make sure no records were lost. Summary output: Create a count by cleaned_status with totals for Received, In Progress, and Done. Add a data quality note listing missing emails, invalid dates, duplicate request IDs, and any unmapped status values.
Known limitations
- The skill cannot analyze data that has not been provided or described with enough detail. - It should not invent rows, columns, values, calculations, trends, correlations, or conclusions. - Data quality issues may limit the reliability of summaries and insights. - Sensitive personal, financial, health, legal, or confidential data should be anonymized or redacted before sharing. - High-stakes analysis may require review by a qualified analyst, domain expert, or responsible decision-maker.