Dataset Profiler

1. Click browse files or drop a CSV, TSV, TXT, JSON, JSONL, NDJSON, XLSX, or XLS file onto the upload area.
2. Review the summary cards for Rows, Columns, Duplicate rows, and Mixed-type columns.
3. Inspect Column, Type, Null rate, Blank rate, Unique, Mean, and Warnings in the schema table.
4. Click Details beside a column to open its type counts, statistics, warnings, and top values.
5. Choose Export JSON, Export CSV, or Export HTML for a copy of the profile.

How Dataset Profiler checks CSV schema and data quality

CSV and TSV files use the first row as headers. An empty header gets a generated name such as Column 1. Each present value is classified as integer, number, boolean, date, or text. The most common type becomes the column type; the column is marked mixed when more than 2% of present values use another type.

Null rate includes null and undefined values plus tokens such as N/A, null, NaN, none, and -. Blank rate counts present empty strings. Unique values leave out missing values. For duplicate rows, the tool trims strings and normalizes missing values before comparing rows.

status,score
yes,10
yes,10
no,N/A

This sample produces Rows: 3, Columns: 2, and Duplicate rows: 1. The status column is boolean. The score column is integer, with one recognized missing value.

How Dataset Profiler reads JSON and spreadsheets

A JSON file can contain a top-level array or a qualifying nested array of records. JSONL and NDJSON files are read one non-empty line at a time, with each line parsed as JSON. XLSX and XLS files use the first worksheet.

The selected file is read in the browser, and this page sends no request containing the file. Invalid JSON, JSONL, NDJSON, or spreadsheet data can show a read error. A file with no rows shows a no-rows error.

Export dataset profile results

Export JSON includes row and column counts, duplicate rows, warnings, type counts, statistics, top values, and per-column warning details. Export CSV includes fields such as column, type, null_rate, blank_rate, unique, numeric summaries, and warnings. Export HTML contains a schema summary with the table shown in the tool.

CSV and HTML exports leave out some drill-down fields, including type counts, top values, quantiles, and text-length statistics. Numeric columns can show minimum, maximum, mean, median, standard deviation, P10, P25, P75, and P90. Date columns show earliest and latest values. Text-like columns show minimum, maximum, and average length.

Where dataset profiling comes up

• A customer export can be checked for empty fields, mixed types, duplicate rows, and columns that look like identifiers.
• A JSONL snapshot can be reviewed line by line, then saved as JSON with its type counts and warnings.
• An XLSX workbook can be checked on its first worksheet for column names, date ranges, unique values, and missing-value rates.

Choose a focused dataset analyzer

Use Dataset Analyze Tools for a broad schema and data-quality pass. Choose Correlation Matrix Generator for pairwise Pearson, Spearman, or Kendall correlations between numeric columns. Choose Fuzzy Duplicate Finder for likely duplicate records based on text similarity, with a choice of record to keep or merge. Choose Time-Series Anomaly Detector for unusual observations in ordered data with timestamps, gaps, and rolling thresholds.

Frequently asked questions