Read-only, always
Data Explorer
The tile that changes nothing. Open the file Excel refused, look at the rows, then look at the columns — what every one of them actually holds, where the nulls and the mixed types and the outliers are, and what is wrong with the file before you build anything on top of it.
Windows 10/11 (x64) · Free · No account
Nothing here writes to your file. Exports and reports land in DataTray's own output folder, and your original is untouched.

What you can do
What people use it for — Data Explorer
The jobs this one actually gets opened for.
- Opening a multi-gigabyte export that Excel will not load at all.
- Working out why a CSV opened as a single column in every other tool.
- Profiling a vendor's file before you trust it — nulls, types, ranges, duplicates.
- Producing a data dictionary for a schema decision you have to justify to someone.
- Running a quick SQL query across a CSV and a Parquet file without installing anything.
What it does
Everything Data Explorer handles
Everything this tool handles, and how it handles it.
Rows, at any file size
The grid pages against the file on disk rather than loading it into memory, so a file with tens of millions of rows scrolls like a small one. Filters and sorts re-run against the file, not against a loaded array.
It shows you its guess
“Detected: semicolon-delimited, UTF-8 with BOM, header on row 1” — stated on screen and every part overridable. The commonest CSV complaint is a file that opened as one column; the fix is to make the guess visible rather than silent.
Profile every column
Inferred type and semantic role — email, phone, date, currency, postcode, ID — plus nulls, distinct counts, min/max/mean/median, quartiles, the top ten values and a distribution sparkline.
A quality score that shows its working
A 0–100 headline over five weighted dimensions, and the weights are printed in the report. It is measured against the file's own shape, never a benchmark, and it says “38% of phone values do not match any phone pattern” rather than “your data is bad”.
A data dictionary you can hand over
One row per column — name, type, semantic role, null percentage, distinct count, range, samples, and a description field you can edit. Exports to .xlsx and to a Frictionless Table Schema .json.
SQL, if you want it
Query files where they lie — SELECT straight FROM a path, join a CSV to a Parquet file, save a view so a path becomes a name. Read-only, with no http:// or s3:// reachable, so SQL pasted off the internet cannot become an exfiltration path.
How it works
From your file to the result
- 1
Drop the file in
CSV, TSV, a fixed-width extract, an Excel workbook, a JSON or NDJSON file, or a whole folder of Parquet.
- 2
Read the rows, then the columns
Start in the grid. When something looks wrong, the Columns tab tells you how wrong and how many rows are affected — and “Show me these rows” jumps straight back to them.
- 3
Take the report
Export exactly what you are looking at — current filter, current columns — or generate the profile report and the data dictionary.
Under the hood
Formats, limits and defaults
Data Explorer never modifies a source file. Its only writes are the exports and reports you ask for.
- Reads
- CSV, TSV, PSV, fixed-width, XLSX, XLS, XLSB, Parquet, JSON, NDJSON, JSONL, XML
- Exports
- CSV, XLSX, Parquet, JSON — whatever is on screen, filters and column selection included
- Reports
- Data Profile & Quality Report (HTML + .xlsx) and the Data Dictionary (.xlsx + Table Schema .json)
- Engine
- An embedded analytical query engine, bundled in the installer — nothing to set up
- Network
- None. Extensions never auto-download and remote filesystems are switched off
Because this tile is interactive, Watch Folders will not propose it and a recipe cannot contain it. Profiling is something you read, not something that should run unattended.
The rest of the belt
The other four tools
Questions
Data Explorer, answered
Get it free — Data Explorer
Data Explorer ships in DataTray, free for Windows 10 and 11. No account, and nothing uploads.
Windows 10/11 (x64) · Free · No account