One category per row, one number each
Ranked barsRevenue by region · votes by party · traffic by source
Length is the easiest visual quantity to compare precisely. Sort by value unless the categories have a natural order.
Upload the workbook. The shape of the data decides the chart, the header row decides the labels, and the sheet comes back with the figure that matters set large enough to read across a room.
Upload a spreadsheetXLSX and CSV · free to try · no formulas required
| Region | Revenue (£m) | YoY |
|---|---|---|
| North | 18.4 | +22% |
| South | 14.1 | +9% |
| Midlands | 11.7 | +14% |
| Scotland | 6.2 | −3% |
| Wales | 4.0 | +6% |
Revenue by region
The North carried the year
Illustrative · source: regions-2026.xlsx
Which chart your columns are asking for
This is the rule the generator follows, written out so you can overrule it when you know something about the data that the data does not say.
Revenue by region · votes by party · traffic by source
Length is the easiest visual quantity to compare precisely. Sort by value unless the categories have a natural order.
Budget split · market share · device mix
Only when the parts genuinely total 100% and there are five or fewer of them. Six slices and it becomes a colour-matching exercise.
Monthly signups · yearly emissions · weekly sales
The slope carries the meaning. Never start the axis above zero on a line meant to show growth — it manufactures a trend.
Plan A vs Plan B · us vs the market · before vs after
Cheaper to read than two charts side by side, because the reader’s eye travels along one row instead of between two pictures.
Progress to goal · capacity used · completion rate
A single percentage does not need a chart at all — it needs to be big, with the target stated beside it.
Product milestones · policy changes · a company history
Space the entries by importance, not by exact elapsed time, unless the gaps themselves are the story.
More on this in the chart maker and in the full chart-choosing guide.
None of this is specific to us — a spreadsheet laid out this way behaves better in every tool that will ever read it, including next year’s you.
Not a title, a logo and three blank rows first. The first row it meets should be the column names.
“Q1 2026 revenue (USD)” is one column. Revenue and margin sharing a column with a slash in between is two.
“Revenue (£m)” beats writing £ in four hundred cells, and it means the chart can label its axis properly.
Merged cells are a layout device. They destroy the row-and-column meaning everything downstream depends on.
A sentence of context alongside the upload — “quarterly revenue by region, 2024 to 2026” — is the difference between a correct headline and a generic one.
Excel workbooks and CSV files, alongside PDFs, Word documents and plain text. If your data lives in Google Sheets, export it as CSV or XLSX first.
No. Uploading raw columns is the intended path — the chart type is chosen from the shape of the data. If you already have a chart in the workbook, it is the underlying numbers that get used, not the picture.
It gets summarised rather than plotted in full. A hundred bars is a wall, not a chart. Expect the top handful ranked, with the remainder grouped — and if you want a different cut, say which column to rank by.
Yes, though results are better when you say which sheet matters. A workbook with a data tab, a working tab and a pivot tab has three different opinions about what the numbers are.
It reads what is in the cells, including the results of your formulas. It is not a substitute for your own modelling — do the maths in the spreadsheet, and let the sheet present the answer.
Yes. Save a brand kit and every piece is drawn in the same palette, so a set of six figures for one report look like they belong to the same report.
Upload the file and get something you would be happy to put in front of a board.
Upload a spreadsheet