data

name	data
title	Tabular Data Analysis
description	Spreadsheet / tabular analysis and charting. Use for any task with tabular reference files (.csv/.xlsx/.parquet/.tsv/.json) or that asks to analyze, aggregate, model, forecast, or visualize data. Wraps pandas, numpy, matplotlib, and scikit-learn for loading, describing, charting, correlation, and quick linear regression.
modalities	["data"]
file_extensions	[".csv",".tsv",".xlsx",".xls",".parquet",".json"]
keywords	["data","dataset","table","csv","excel","spreadsheet","dataframe","analysis","analytics","statistics","aggregate","pivot","chart","plot","graph","histogram","correlation","regression","forecast","trend","kpi","metric","model","predict"]
requires	["pandas","numpy","matplotlib","scikit-learn"]
version	1.0.0

Data Skill — analyze & visualize tables

Use this skill to load tabular reference files robustly and turn them into analysis + charts for the deliverable.

A typical "data problem" workflow:

1. df = data.read_table(path) — load CSV/XLSX/Parquet/JSON into a DataFrame.
2. summary = data.describe(df) — shape, dtypes, numeric stats, nulls.
3. data.quick_chart(df, x, y, kind="bar", out="chart.png") — a shippable PNG.
4. (optional) data.correlation(df) / data.linreg(df, x, y) for modelling.

Toolkit API

from skills import data

data.read_table(path, **kw) -> pandas.DataFrame   # dispatch by extension
data.describe(df) -> dict        # {shape, columns, dtypes, numeric, null_counts}
data.quick_chart(df, x, y=None, kind="line", out="chart.png", title=None) -> str
data.correlation(df, out=None) -> dict   # numeric correlation matrix (+ optional PNG)
data.linreg(df, x, y) -> dict    # {slope, intercept, r2, predict_fn_repr}

Notes:

• read_table always returns a real pandas.DataFrame — inspect df.columns/df.dtypes before assuming names (never hardcode columns).
• Charts render headless (Agg backend); always out= to a file.