csv.design

Prepare your CSV

Prepare sanitized sample data.

Fix your sample once and reuse it every month.

What the app accepts

File up to 2 MB
Up to 50 columns
First 1,000
rows used
invoice_idcustomeramountstatus…
INV-001Fictional Customer 11200Paid…
INV-002Fictional Customer 2980Open…
…
  1. Headers in row 1, up to 120 characters each.
  2. Text cells: the first 80 characters are used. Category labels are kept up to 500.
  3. One table, comma, semicolon, tab or pipe separated. UTF-8 works best.

The rules

Column names

They label your charts and match your real CSV to the template.

Do

  • Short and unique, such as invoice_date.
  • Identical every export, capital letters included.

Avoid

  • Line breaks or tabs in a header.
  • Units or notes, such as Amount (USD).

Categories

The design only knows the values it can see.

Do

  • Include every category at least once, the rare ones too (such as Disputed).
  • Spell each value one way.
  • Keep yes/no columns to one pair: true/false, yes/no or y/n.

Avoid

  • An ID, a name or free text as a category.
  • Leaving a category out because it is only a few rows.

Numbers

One stray word turns the whole column into text.

Do

  • Plain digits, such as 1200.50.
  • Leave empty cells empty.

Avoid

  • N/A, - or TBC in a number column.
  • Units in the cell, such as 12 kg.

Dates

Dates drive trend charts.

Do

  • One format per column, ideally 2026-01-31.
  • Enough periods for a trend, such as six months.

Avoid

  • TBC or next week in a date column.
  • Ambiguous slash dates such as 04/05/2026.

Text and titles

Free text shows in tables and record views.

Do

  • Keep values short and on one line.
  • Leave out long comment columns.

Avoid

  • Line breaks inside a cell.
  • Paragraphs: only the first 80 characters are used.

Rows

Enough to show every pattern, and no more.

Do

  • Cover every category and period at least once.
  • Stay within 1,000 rows.

Avoid

  • Totals or subtotals rows.
  • Thousands of near-identical rows.

Sanitized sample data

Same columns as your real file, with made-up values or anything sensitive removed. Your real CSV stays on your computer.

Do

  • Replace names, emails and IDs with made-up ones.
  • Keep realistic ranges and categories.

Avoid

  • Real personal or financial details you have not checked.
  • Passwords or keys in any column.

Spreadsheet export

Most problems come from the export, not the data.

Do

  • Excel: Save As, CSV UTF-8. Google Sheets: File, Download, CSV.
  • One worksheet, with no title rows, merged cells or subtotals.

Avoid

  • Several tables on one sheet.

When files differ

The template matches your real CSV by column name.

Do

  • Extra columns are ignored.
  • A missing column leaves its cards blank.
  • A wrong kind of value is flagged.