Skill

Exploratory Data Analysis

k-dense

Data

Profiles a dataset you already have - distributions, missingness, outliers, leakage, and transformation sensitivity - and scaffolds a rigorous EDA report, all computed locally and deterministically without touching the network. It INSPECTS data before you model it; it does not gather the data, draw the finished charts, or make causal or confirmatory claims.

Situations it fits

You inherited a messy spreadsheet of customer purchases and need a careful, offline pass to find gaps, weird numeric values, and any columns that might accidentally reveal the label before you train a model.

Exploratory Data Analysis. Exploratory-data-analysis (k-dense) is the right fit because the task is to examine a dataset you already have: locate anomalies, missing entries, and potential leakage, and recommend transformations before modeling. It focuses on local, deterministic inspection and reporting about the table itself. The other skills either produce visual outputs, fetch web content, or synthesize multi-site research rather than performing this pre-model data check.

You’ve got a single CSV of past customer purchases and need a quick pass to find surprising values, which fields are unreliable, and what needs cleaning before any modeling work starts.

Exploratory Data Analysis. Exploratory data analysis inspects a dataset you already have to identify odd values, missing or unreliable fields, and other issues to fix before modeling; it’s the right first step here. The other skills either fetch or synthesize information (web-scraping and deep research) or turn an already-prepared dataset into visual output (chart design), so they don’t replace the initial local inspection this scenario requires.

You handed a 2 GB CSV of user events to a teammate and they asked for a short report listing data completeness, weird fields, and any columns that would break a predictive model.

Exploratory Data Analysis. exploratory-data-analysis inspects an existing dataset to summarize missing values, odd entries, and other quality issues that matter before modeling. It’s the right fit for producing a concise diagnostics report. The other skills are for visualization, fetching external content, or broad web synthesis and don’t specialize in local dataset quality checks.

6 scenarios in the bank answer to Exploratory Data Analysis. The rest are in the game.

Skills it gets confused with

These share a family with Exploratory Data Analysis, which is another way of saying they are the ones you might reach for by mistake.

Knowing what Exploratory Data Analysis does is the easy half. Telling it apart from the others under time pressure is the game.

Today's session

Exploratory Data Analysis is part of k-dense-ai/scientific-agent-skills. Licence: MIT. The description above was written for this game, not taken from the skill.