Rows and columns
Confirm the file has enough rows, clear headers, and columns that match the intended use.
CSV data readiness before analysis or AutoML
Data readiness means the CSV is understandable enough to review, report on, share, clean up, or prepare for model training.
MLdeck Data Quality gives browser-local readiness signals without starting training or uploading the CSV for analysis.
What to check first
Missingness
Review blank cells and decide whether they are expected, accidental, or important to explain.
Schema clarity
Look for mixed formats, confusing names, dates, IDs, and fields that need domain context.
Risk indicators
Inspect identifier-like columns, constant columns, high-cardinality fields, and other review signals.
How to use readiness signals
A readiness score is a summary, not a verdict. A clean-looking score can still miss domain issues, and a lower score may simply mean the file needs documentation or cleanup before the next step.
Use readiness signals to slow down at the right moments: before a report is shared, before a CSV is sent to a teammate, before cleanup assumptions are made, and before AutoML metrics are interpreted.
For non-ML and ML users
Business users can use Data Quality to decide whether a spreadsheet is understandable enough to discuss. Analysts can use it to prioritize cleanup. ML users can use it to find early warnings before selecting a target and training models.
The shared point is the same: review the CSV before decisions depend on it.
Continue the review
After readiness review, decide whether the next step is documentation, cleanup, sharing, or model training. Data Quality is useful before AutoML, but it is not a model-training workflow.