Paste a CSV dataset into "Dataset Input" or upload a CSV/JSON file. The profiler inspects each column and gives you a quick quality snapshot before the data moves into BI, ETL, or machine-learning steps.
What the tool checks:
- Missing values per column
- Duplicate rows, or duplicate combinations based on the columns you list in "Duplicate Key Columns"
- Column type inference: number, boolean, date, string, or empty
- Numeric outliers using an IQR-style rule
- Format drift for string/date-like columns, such as mixed date styles or code-vs-free-text inconsistencies
How to fill the fields:
- Dataset Input: paste CSV text directly when you want a quick profile
- Dataset File: upload CSV or JSON if the dataset is larger or already saved locally
- Duplicate Key Columns: optional comma-separated keys such as id,email to detect duplicates by business key instead of whole-row matching
- Sample Rows: controls how many example rows appear in the report preview
How to read the report:
- Quality score is a simple 0-100 summary where more missing cells, duplicate rows, and anomaly signals reduce the score
- Missing shows how many blank/null cells were found in that column
- Distinct shows how many unique values appear in the sampled dataset
- Anomalies highlights numeric outliers
- Format drift highlights columns where values look structurally inconsistent
Current scope:
- CSV and JSON are supported
- JSON should be an array of objects or an object with a rows array
- The score is meant as a quick operational signal, not a formal data-governance grade