Start with a column-by-column profile
Upload a CSV file to see its rows, columns, detected data types, and missing values. Use this first pass to find incomplete fields and check whether numeric and categorical columns have been interpreted as expected. Review parser warnings before relying on the results.
Look beyond averages
Distributions reveal skew and unusual values that a single average can hide. Compare numeric columns, inspect category frequencies, and use correlation heatmaps to look for relationships worth investigating. A correlation is a starting point for analysis; it does not establish cause and effect.
Example: inspect a customer churn export
For a file with tenure, monthly charges, contract type, and churn, start by checking missing cells and data types. Then compare tenure and charges distributions and examine relationships among numeric fields. Treat the product preview as an illustration, and use the demo to explore a sample dataset before uploading your own.
Move from exploration to a reusable result
Review suggested cleaning operations, apply the ones appropriate for your dataset, and continue into feature engineering. Export a cleaned dataset, notebook, or report so you can take the result into your next analysis.
How to get started
- Upload your CSV and review parsing warnings.
- Inspect column quality, distributions, and correlations.
- Review cleaning suggestions and export your results.