Tutorials
Three short guides that cover most of what people use SheetSense for. Each one uses the bundled sample dataset, so you can follow along without your own file.
- Guide 1 · 2 minutesTranslate any formulaUse this when you open a workbook and find a formula you didn't write.
- Guide 2 · 3 minutesClean a column without formulasUse this for exports where the same thing is written five different ways: phone numbers, dates, names, region spellings.
- Guide 3 · 2 minutesAsk your dataUse this for questions you'd normally answer with a pivot table: who sold the most, which product leads a region, where the blanks are.
Guide 1 · 2 minutes
Translate any formula
Use this when you open a workbook and find a formula you didn't write. The example is a commission formula from a forecast file someone else built.
- 01
Copy the formula from the formula bar
Click the cell, then copy from the formula bar at the top. Copying the cell itself gives you the result (10,120.00) instead of the formula. Google Sheets works the same way.
- 02
Paste it into the translator
The public translator needs no account and allows 5 a day. The translator inside the workspace uses a larger model, which handles very long nested formulas better, and counts as one AI action.
- 03
Read the Plain English tab against your sheet
The steps run from the inside of the formula outward and name the real cells. Check each one against the sheet: if a step says “C14 is deal status” and column C actually holds the close date, the formula is doing something different from what its author intended.
- 04
Switch tabs for the version you need
The Excel tab sometimes rewrites old patterns with newer functions, like nested IFs as IFS. SQL assumes a table called
dataand pandas assumes a DataFrame calleddf; rename them to match yours. - 05
Read Watch out for before you trust the number
This list covers the inputs that quietly break the formula: blanks that aren't really blank, numbers stored as text, duplicate lookup keys. On Pro, you can publish the translation as a public page and link it from the workbook.
Guide 2 · 3 minutes
Clean a column without formulas
Use this for exports where the same thing is written five different ways: phone numbers, dates, names, region spellings. The sample dataset has all four, so it's the best place to practice.
- 01
Load a sheet
Upload a CSV or Excel file from the Sheets page, or click Load sample dataset for 200 messy orders. Files are parsed in your browser; if one has more than 5,000 rows, you'll see how much will be kept before anything is saved.
- 02
Open Clean and pick the column
The column you pick turns green in the grid so you can see what you're about to change. The small tag in each header (Aa, 123, date, tel) shows what most of the column contains, which helps spot a phone column stored as numbers.
- 03
Choose where the results go
Same column rewrites values in place. New column keeps the original next to the cleaned version. Split breaks one column into two to four new ones, like Customer into First name and Last name.
- 04
Describe the change the way you'd tell a coworker
Say what to do with the awkward rows: “Format as (XXX) XXX-XXXX. Drop a leading +1. Leave blanks empty.” Instructions that cover edge cases come back more consistent than “fix the phones”.
- 05
Check the diff, then apply
Changed cells are amber with the old value struck through; new columns are green. Tick Show unchanged to see rows that were left alone. Apply writes the change to your saved sheet, and Undo appears right after in case you change your mind. Download CSV takes the cleaned sheet back to Excel.
Guide 3 · 2 minutes
Ask your data
Use this for questions you'd normally answer with a pivot table: who sold the most, which product leads a region, where the blanks are.
- 01
Name the columns in your question
“Which Sales rep has the highest average Revenue per order?” gets a better answer than “who's the best rep?”, because there's no guessing about which column means “best”.
- 02
Read the answer, then the reasoning
The reasoning says how the answer was reached, including rows it skipped. If it grouped or filtered differently from what you meant, rephrase and ask again.
- 03
Confirm it with the formula on big sheets
On sheets with more than 200 rows, Ask reads an even sample of 200 rows from across the sheet. A sample is good for spotting which rep or product leads. For an exact total, paste the formula it gives you into your spreadsheet, or use the pivot table setup.
Try it on the sample dataset
A free account comes with 10 AI actions, enough to run every guide above.