Checklist

Exploratory Data Analysis — EDA Quick Checklist

An actionable, saveable EDA checklist that captures dataset provenance, quality checks, visualizations, candidate hypotheses, and recommended next steps so teams reliably produce discovery-ready artifacts and feed the discovery backlog.

Interactive Tool

Exploratory Data Analysis Quick Checklist

Use this checklist to rapidly profile a dataset, record key findings, and produce discovery-ready artifacts. Capture exact dataset provenance, basic quality checks, visual signals, candidate hypotheses, and a recommended next step so your EDA leads to prioritized experiments instead of ad-hoc reports.

Record the exact version, file path, database table, query, and date range used (e.g., prod.users_2026-08-01.sql).
Where did the data come from? upstream system, extraction method, refresh cadence, known transforms.
Checked overall and by key segments (percent missing, patterns by time or group).
Note any columns with high missingness or segment-specific gaps and whether imputation/filters are needed.
Used visual checks and/or robust statistics to find extreme values, recording suspected data issues vs true signals.
Describe notable outliers, potential causes, and whether to filter or investigate further.
Check distributions for numeric variables and by segments such as region, cohort, or customer type.
List the segmentation variables you inspected (e.g., country, plan_type, acquisition_channel).
Look for strong relationships, confounders, and cohort-based shifts; note methods used (Pearson, Spearman, cohort aggregates).
Summarize any strong or surprising relationships and potential confounders.
Record testable hypotheses that follow from observed signals, along with an evidence confidence level.
For each hypothesis include: short statement, why it matters, supporting evidence, and confidence (low/medium/high).
Ideally a notebook or script that can re-run the checks and reproduce figures/tables.
Provide link to the notebook, repo, or storage location (optional but recommended).
Exported aggregates, cohort tables, and figure assets that can be attached to backlog items or experiments.
Link to the backlog item, experiment ticket, or shared folder where follow-up work will be tracked.
Quickly indicate whether the recommended follow-up should be High, Medium, Low, or Deferred.
E.g., 'Run A/B test on X; rationale: correlation present across cohorts, medium confidence.'
Anything else useful (data owner to contact, suspected data pipeline issues, quick visual references).
You can explore this tool now. Sign in or create an account to save your responses and return to them later.
Sign In to Save Create an Account
Make this tool part of your work

Save a personal copy, bring it to your team, or tailor the questions and workflow to fit what you are hungry to improve.

Member customization and team collaboration are coming soon.
Discuss Ask the Librarian

Discussion

Comments and conversation will live here.