SigmaForge
Your data

Four ways in. No integration project.

Paste it, upload a spreadsheet, drop a screenshot of a table, or use the measurements your team already recorded on the project.

Try the SPC demo
4
ways in
and no others
28/32
deterministic
figures from the engine
32
analyses
on the same data
22
chart types
drawn from it
DataForge yield chain for a server mainboard burn-in process: first-pass yield by step, rolled throughput yield of 89.48%, and the burn-in cycle flagged as 55% of the total loss.
Product screen, demo data

Four ways in

Pick whichever is fastest for the data you have.

Every DataForge data field offers the same routes. All four end in the same place: the numbers in the field, where you can see and edit them before anything runs.
  1. 01

    Paste values

    Type or paste a column of numbers straight into the tool.

    Typed or pasted into the field

    Commas, spaces, tabs or a column copied out of Excel all work.

  2. 02

    Upload Excel or CSV

    The spreadsheet is parsed in your browser, so the file never has to leave it to be read.

    Read in your browser

    Excel (.xlsx, .xls, .xlsm), CSV or TSV. The first sheet lands in the same field you would have typed into.

  3. 03

    Drop a screenshot

    A photo or screen capture of a table is read back as rows you can check before you use them.

    Sent to be read, then handed back

    A photo of a logsheet works. The numbers come back as rows you look over before anything runs.

  4. 04

    Use project measurements

    Pull the readings your team already recorded on the project collection sheet.

    From the project collection sheet

    Pick the process step; the readings your team recorded there fill the field, with where they came from noted.

Before it runs

You check what was read, not a format.

Under every data field is a live readback of what the tool understood: how many values, the mean, and any gaps. Whether the numbers were read as percentages or raw counts is stated rather than guessed, with a one-click correction beside it. Then the analysis runs on numbers you have already looked at.

  • How many values were read, the mean and any gaps
  • Percentages or raw counts, stated and correctable
  • Where the numbers came from, labelled on the field
sigmaforge.ai/dataforge
DataForge yield chain for a server mainboard burn-in process: first-pass yield by step, rolled throughput yield of 89.48%, and the burn-in cycle flagged as 55% of the total loss.
  1. 1Saved to the projectWith the date, and read-only once saved.
  2. 2The biggest loss, on the chartBurn-in, 55% of everything lost.
  3. 3The figuresRolled throughput yield 89.48%, from the engine.
Product screen, demo data

The engine

Your data goes to the engine first.

Of the 32 analyses, 28 compute every figure in a deterministic engine. The model writes the explanation over the top, and only that.

  • The engine

    Means, limits, capability, test statistics, p-values, yields.

  • The model

    The plain-language reading of those figures, and what to do next.

28of 32computed by the engine

On the same data

32 analyses, filed by phase and belt.

Once the numbers are in, every analysis in the library can use them: capability, hypothesis tests, ANOVA, regression, control charts and the rest, each with a preview of its chart, the DMAIC phase it serves and the belt that teaches it.

  • 22 chart types
  • A test selector when you are not sure which one
  • Every result saved into the project
See DataForge
sigmaforge.ai/dataforge
The DataForge analysis library: every analysis as a card with a preview of its chart, its DMAIC phase and its belt level, filterable by phase and belt.
Product screen, demo dataThe DataForge analysis library

What we do not connect to

Four routes is the whole list.

We would rather say what is not there than let a diagram suggest it. If your data lives in another system today, export it and use route two.

  • No ERP or MES integration

    Nothing reads from your plant or business systems directly.

  • No connectors

    There is no live feed from machines, historians or databases.

  • No webhooks

    The platform does not push events to other systems.

  • No public API

    There is no programmatic way in or out today.

Export, then upload.A spreadsheet out of any system goes in through route two, read in your browser.Try it with a CSV

Questions

Where your data goes.

  • Is my spreadsheet uploaded to your servers?

    The file itself is not. It is read in your browser (the first sheet of a workbook) and its values are written into the data field, where you can see and edit them. What is in that field is sent when you run the analysis, and nothing before.

  • What happens to a screenshot?

    A screenshot is the one route that leaves your browser as a file: the image is sent to be read, and the table in it comes back as rows. You look over those rows before anything is calculated.

  • Which file types can I upload?

    Excel (.xlsx, .xls and .xlsm), CSV and TSV, and images of a table.

  • Can it pull from the systems we already run?

    Not directly. There are no connectors today. Export what you need as a spreadsheet and upload it, or paste the column in. Both take seconds.

  • What are project measurements?

    Readings your team recorded on a project’s collection sheet, by process step. Pick the step and its readings fill the field, labelled with where they came from.

  • Can I try it without an account?

    Yes. The live control chart and the capability calculator both run in your browser, with no account and nothing to install.

More in platform

No account needed

Drop in a CSV and watch it work.

The live control chart runs in your browser. When you want the other analyses, the documents and the project, book thirty minutes with us.

Try the live chartBook a demo