Updated
From data to decision: agent skills for analysts
From data to decision: agent skills for analysts
Analysts spend most of their time getting data into shape before the actual analysis starts. Skills flip that ratio: the cleaning and reshaping become repeatable, and the analyst's time goes to the decision instead.
Spreadsheet building
A spreadsheet skill turns a plain-language request into a working workbook — formulas, pivots, and charts included. Instead of hand-building a model from scratch, you describe what the sheet should show and get the structure back, ready to review.
Exploratory data analysis
EDA is the step where you find out what a dataset actually says before you commit to a story. An EDA skill runs the standard checks — distributions, missing values, outliers, correlations — and reports them in order, so nothing obvious is skipped.
Chart and report production
Once the numbers are clean, a charting skill produces the visuals and the written summary in the same pass. The chart and the takeaway stay in sync because the same skill generates both.
A typical analyst stack
Put three skills together and most reporting work becomes one pass:
- Spreadsheet builder for the workbook and the model.
- EDA for the first look at what the data says.
- Charting for the visuals and the written takeaway.
The skills share a workflow, so the output of one feeds the next without reformatting in between.
Why these work for analysts
Data work is a good fit for skills because it's procedural: the same steps run on every new dataset, and the only thing that changes is the input. A skill encodes those steps once and applies them consistently, which is exactly the discipline that ad-hoc prompts tend to lose on the tenth run.
Where to go next
The catalog's data analyst skills cover these patterns and more. Pair them with a spreadsheet or charting skill to go from raw file to finished report in one pass.