Common data issues and how to avoid them

The most common problems

Most chart issues are caused by a small number of avoidable data mistakes.

Duplicate header names

If two columns use the same name, the sheet becomes invalid.

Fix: Rename columns clearly and specifically.

Numbers stored as text

It is not enough for a value to look like a number. It must be a numeric cell so that aggregation can be performed. An exception is values in a table, since aggregation is not performed there.

Bad

• "7200"

Good

• 7200

Fix: Check formatting and remove apostrophes or text-only formatting.

Totals, notes, or helper rows in the dataset

The backend imports every row after the header as data.

So rows such as these will cause problems:

• Total

• Notes

• Source: internal analysis

• Units: EUR

Fix: Keep these outside the dataset tab or remove them entirely.

Mixed values in one column

A metric column should not contain entries like:

• 120

• n/a

• -

• unknown

Fix: Leave the cell empty if a value is missing, or correct the source data.

Empty or unstable labels

Dimension labels should stay clear and consistent.

Avoid

• Berlin

• berlin

• BERLIN

• Berlin

These may be treated as different values depending on usage.

Fix: Use one consistent naming style.

Unsorted data

Since there is no automatic sorting, charts may appear in an unexpected order.

Fix: Sort the dataset in Google Sheets before publishing.

Good vs bad example

Good example

Why this works:

• one clear header row

• all headers are unique

• dimensions and metrics are easy to identify

• number columns contain real numbers

• no helper rows or merged cells

Bad example

Why this fails:

• duplicate headers

• mixed content types

• note-like content inside the dataset

• totals row included as if it were data

Final publishing checklist

Before handing over or using a Google Sheet for a chart, check:

• the first row contains the headers

• every header is unique

• every following row is real data

• dimensions are text labels

• metrics are numeric cells

• filters use string values where needed

• chart-specific requirements are fulfilled

• the sheet is sorted correctly

• labels are stable and consistent