● Your data, ready for what comes next

Make your data model-ready.

Find data quality issues, review smart cleaning suggestions, and understand how prepared your dataset is for machine learning — before you build a model.

↥

Start with your dataset

Choose a supported tabular file to inspect its quality.

▤

Upload your data fileCSV, XLSX, JSON, Parquet, Arrow IPC or Feather · up to 50 MB

Files are sent to your configured NUMPA API for in-memory processing. The MVP does not intentionally persist uploaded datasets on the server.

◈

ML readiness assessment

Choose the intended ML task so the assessment can use relevant checks.

Intended ML task

Target column name

✳

Smart cleaning controls

Choose the transformations you want to preview. The source file remains unchanged.

Fill missing values
Use safe column-based defaults from the cleaning engine.Remove exact duplicate rows
Keep this off if duplicate records may be meaningful.Normalize text whitespace
Trim leading and trailing spaces in text values.

Review the report before using cleaned data in a real project. Automated cleaning cannot know every domain-specific rule.

▦

Dataset overview

A quick read on structure and quality before any changes.

▤

Your data story starts here

Upload a dataset and select “Inspect dataset” to see its size, missing values, duplicate signals, and profile.

◉

Readiness score

Understand strengths and quality gaps for your chosen ML task.

◉

Know before you train

Choose a task and target column, then calculate the readiness score to see the factors affecting your dataset.

✳

Cleaning preview

Check the proposed output before downloading the cleaned CSV.

✳

Nothing changes without review

Select your cleaning options and preview the report. The original upload is never overwritten by this workflow.