DATA & ANALYTICS / DATA HYGIENE

Data Hygiene Profiler

Clean and standardize CSV values, review changes and export the result.

Give your data a clean start.

No file loaded

Recipe Builder

Structural Repairs

Sanitize Headers Trim spaces, replace spaces with underscores, and deduplicate identical column names. Prune Empty Rows Remove rows where every single cell is completely blank. Strip Control Characters Remove ASCII control characters, zero-width spaces (U+200B), and BOM markers (U+FEFF). Preserves language and emoji joiners (U+200C/U+200D).

Formatting & Cleansing

Trim All Whitespace Remove leading/trailing spaces across the entire dataset.

Reformat & Standardize Dates

Input order Output format

Ambiguous numeric dates use Input order. Output format controls the exported date; ISO and named-month inputs do not depend on numeric order.

Clean Numeric / Currency

Strips currency symbols, fixes 'k' (e.g. 60k), and parses word numbers.

Standardize Phone Numbers

Converts raw phone numbers to standard XXX-XXX-XXXX format.

Title-Case Text

Properly capitalizes names and addresses (e.g. helen keller → Helen Keller).

Missing Data Handling

Treat these values as missingAdditional markers (one per line)Matching ignores case and surrounding spaces. Unchecked markers remain valid text. Exports preserve stored values.

Drop Rows Missing Critical Data In:

Impute (Fill) Missing Data

Deduplication

Deduplicate Based On:

Workspace & Preview

Upload Dataset (CSV)

Drop your file here or click to browse. Processes securely in your browser RAM.

Ready to load a CSV file.

A CLEARER WAY TO CLEAN CSV DATA

From messy rows
to a reliable starting point.

Data Hygiene brings CSV cleaning, inspection and reusable recipes into one browser workspace. Check the source, explore changes, and review the result before exporting.

How to clean a CSV

Load and inspect. Open a CSV and inspect the original values. Use issue filters to find missing values, surrounding whitespace or validation failures.
Choose your rules. Enable the repairs you need, choose formatting columns, set your missing-value policy and add fills or deduplication. Structural and whitespace repairs start unchecked.
Review the proposed changes. Switch to Preview with changes or Source / preview comparison. The live preview processes the full dataset while preserving the source and completed output.
Execute and export. Run the recipe, inspect the completed result, then download the cleaned CSV. Save a named preset or export recipe JSON for the next dataset.

Tools for the details that matter

Clean and standardize

Sanitize headers, prune empty rows, strip control characters and trim whitespace. Format dates, numeric and currency values, US phone numbers and title-case text in selected columns.

Find and explain issues

Inspect affected rows and columns, separate missing values from whitespace issues, and apply validation rules. Compare source and changed values with retained source row identities.

Repeat a considered recipe

Save presets on this browser, share settings through recipe JSON, and inspect a concise Python-style outline. Applying settings updates the preview; Execute creates the completed export.

A FEW THINGS WORTH KNOWING

Clear answers.
Confident next steps.

How your data is handled, what the controls mean, and how to keep your work.

Back to the workspace ↑

How do source, live preview and completed output differ?

Original source shows untouched input. Preview with changes applies the current recipe to the full dataset, with Source / preview comparison available. Execute creates completed output for Download CSV. Later preview edits do not overwrite that export. Export affected rows instead uses the selected review view and issue type, including all matching rows and all columns across pages.

Can I control missing values and how they are filled?

Toggle blank cells, N/A, NULL, NaN and the dash marker, or add custom markers one per line. Matching ignores case and surrounding spaces. Unchecked markers remain text. Fill with mean, median, mode or custom text; literal replacements such as N/A are preserved. Numeric fills use usable values across the full dataset, excluding configured missing markers and empty numeric strings. A numeric fill cannot run without usable numbers.

How do deduplication and validation rules work?

Exact deduplication compares all fields; key deduplication compares selected columns. The first matching row stays, and comparisons retain source row identities for removed records. Validation flags values without modifying them: required values, inclusive numeric/date ranges and exact allowed categories. Date checks expect valid YYYY-MM-DD values. Optional values classified as missing skip range and category checks.

Does my data leave the device, and how large can it be?

CSV parsing, cleaning and review run in browser memory; uploaded CSV contents are not sent to a processing server. Website assets and third-party libraries can require network requests, so offline availability is not guaranteed. Practical capacity depends on browser/device memory, data and recipe. Background workers help with cleaning, but large previews and comparisons may take time. Keep the original CSV and download results before closing.

How do I reuse a recipe, and is the Python preview executable?

Named presets are stored in this browser, including column names, replacement text, missing markers and validation settings, not dataset row arrays. Clearing browser storage removes them; export recipe JSON to keep a backup or move settings. The Python preview is an illustrative pipeline outline with helper calls, not a standalone executable. Close preview or press Escape to return without losing your dataset.