Spreadsheets are the right shape for data work. A sheet you can edit, formulas and plots that update live, every value visible in a cell. That's why analysts, scientists, and anyone working with numbers keep coming back to them.
The shape is right. The feature set is incomplete. Three things that are table stakes in any real programming environment never made it into the spreadsheet: layered plotting, Monte Carlo simulation, and direct warehouse queries. I'm a data scientist and engineer, and I feel the gap every time I leave the spreadsheet for R or Python to do work that should be doable in the spreadsheet.
TukeySheets is the spreadsheet with those three things added.
1. Powerful plotting built in
TukeySheets gives you a layered, ggplot-quality plot that sits next to your data and updates as you edit it.
Ggplot itself is excellent: layered grammar of graphics, sensible defaults, beautiful output. But to use it, you leave the spreadsheet, load data into R, and re-run the script every time anything changes. The plot is always one step behind the truth. In TukeySheets, you get beautiful plots, and edit a cell and the plot updates. No re-running, no re-loading, no script.
2. Monte Carlo as a first-class primitive
TukeySheets treats stochastic columns as variables: stable, referenceable, redrawn on demand.
Spreadsheets have always had random functions, but they're volatile. Every recalculation re-rolls everything at once, so you can't hold one column steady while you re-roll another. The workaround has been to leave the spreadsheet entirely. In TukeySheets, =SAMPLE.NORMAL(0, 1, 10000) spills ten thousand draws down a column and pins them. Click the resample icon to redraw just that column, and everything downstream updates with it. Monte Carlo isn't a bolt-on; it's how the cells work.
3. Query a warehouse, save the query with the data
TukeySheets writes SQL against your data warehouse, lands the rows in a sheet, and saves the query alongside the data it produced.
You shouldn't have to export a query result to CSV just to look at it. You should be able to iterate on the query until it's right, then hand the file to a colleague who can rerun it against today's numbers with one click. That isn't easy to do in R or Python either. In TukeySheets, it's the default.
A native app, not a tab
TukeySheets is a native desktop app: your data stays on your machine and the spreadsheet stays fast. It doesn't live in a browser tab, it doesn't ship your data to someone else's server, and it doesn't spend a ton of resources doing web things. Performance comes from running close to the metal; privacy comes from your data not leaving.
What about AI?
The features above stand on their own, but they also make the spreadsheet a surface an AI can drive without losing the readability that makes it trustworthy.
The spreadsheet has always been the audit surface for data work: a place where every intermediate value sits in a cell you can inspect, tweak, and reason about. Adding plotting, simulation, and warehouse queries turns that surface into one an AI can actually drive, with every step still legible to a human. That's a longer post. Soon.
Try it
Open beta begins June 2026. Sign up to be notified.