SPSS Tabulator: crosstabs with significance from a .sav file

Upload an SPSS .sav file and get back a professional crosstab workbook in Excel, the kind market researchers build by hand in SPSS, Q or Displayr:

• banner × stub tables with column-letter significance tests (90 / 95 / 99 %), with a Bonferroni correction across the columns of each banner block, so a letter is stricter than a plain two-proportion test
• nets (Top 2 Box, Bottom 2 Box, your own groupings)
• multi-response and grid questions grouped automatically
• weighting, with effective base (Kish) for the tests
• means with t-tests on scale questions

No SPSS license needed. The tables come from a deterministic statistics engine: the same file and the same spec always give the same numbers. By default no AI is involved at all. There is one optional extra, an AI-written executive summary sheet (automatic mode only, off unless you turn it on); when enabled, the aggregated tables, not the raw respondent rows, are sent to an AI model, and its text can contain mistakes. The tables stay exact either way.

Try it free in one click

The input form starts with a synthetic sample survey (400 respondents: gender, age, region, a satisfaction scale, a multi-response brand question and a rating grid). Press Start and you get the full automatic workbook, with banners, nets and significance letters. Runs on the sample file are never charged. Then upload your own .sav.

Three ways to run it

Mode	What you give	What you get	Price
Automatic workbook	just the .sav	the engine picks banners, groups multi-response and grids, adds nets, builds the full book	per workbook, see below
Workbook from my spec	the .sav + a JSON spec	exactly the tables you asked for	per workbook, see below
Variables and labels	just the .sav	variable names, labels, value labels and detected question types	free

Tip: run Variables and labels first to see your variable names, then write a spec.

Spec example

{
  "banner": "GENDER",
  "stubs": ["Q1", "Q2", "Q3"],
  "weight": "WEIGHT",
  "nets": { "Q1": { "Top 2 Box": [4, 5], "Bottom 2 Box": [1, 2] } },
  "title": "Brand tracker - wave 3"
}

Pricing

You pay once per delivered workbook, based on the number of respondents (cases) in your file. If the run fails, whether because of a bad file, a rejected spec or an error on our side, you are not charged.

Respondents in the file	Price per workbook
up to 1,000	$3
1,001 – 5,000	$6
5,001 – 20,000	$12
more than 20,000	$20

Output

• The workbook, as an .xlsx file in the run's key-value store (record OUTPUT), with a download link in the dataset.
• A dataset row with the respondent count, the amount charged, the engine version, the run's statistics and the assumptions the engine made (which variables it grouped, which scales it detected, what happened to each column), so you can audit the book.

Limits

• .sav files only, up to 50 MB.
• Workbook language: English by default, or Spanish with Workbook language. It only changes the words the workbook writes by itself (Nets, Mean, Summary, Significance, Base...). Your variable names, question texts and answer labels always appear exactly as they are in your file.
• Very wide files (over ~1,000 variables) with many banners can hit the processing time limit; split the spec if that happens.

What to check in automatic mode

Automatic mode is a fast starting point, not a finished deliverable. On a real survey it picks sensible banners (for example city, gender), groups multi-response questions and adds nets and significance letters. But it decides on its own, so review the assumptions in the dataset row and the first sheet of the workbook. Known limits:

• It does not detect a weighting variable. If your study is weighted, use Workbook from my spec and set weight.
• Administrative columns (respondent ID, timestamps) can be tabulated as if they were questions.
• It chooses the banners for you. If you need specific cross-breaks, use Workbook from my spec.

Your data

Your file is downloaded by the Actor, read by the tabulation service to build the workbook, and the result is stored in your own Apify account. The service processes each file to answer that request and does not add it to a file library. The workbook is yours to delete from your run storage at any time. Do not upload data you are not allowed to process outside your organisation.

Automate every survey wave (n8n, Make or your code)

Tracking studies are where this saves the most time: the same tables, every wave. A typical n8n workflow watches a Google Drive folder, uploads each new .sav to your Apify storage, runs this Actor with your study's spec, saves the workbook to another folder and emails the analyst.

To call it from anything else, use the Apify API endpoint run-sync-get-dataset-items with savFile set to a URL of your file. The dataset row includes workbookUrl, a signed link to download the Excel. The run fails without charge if the workbook is not produced.