Lotus Improv Helped Define the Spreadsheet Pivot

Lotus Improv is a key early program behind modern spreadsheet pivot tables. Lotus released it for the NeXTSTEP platform in 1991, before its Windows version appeared in 1993. Its rearrangeable data views provided the model for modern spreadsheet pivots. (Release history, pivot-table history)

That is more precise than calling it the first program to summarize data by category. In his account of Improv’s origins, Pito Salas acknowledges that cross-tabulations already existed in statistical software. He describes Improv’s innovation as “flexible views”: changing which dimensions appear across or down a table while keeping the underlying data and formulas intact. (Salas’s account)

What Improv changed

Improv separated data, formulas and views instead of putting everything inside spreadsheet cells. A formula could use meaningful names such as Profit=Revenue-Expenses, rather than cell addresses.

That separation made it possible to inspect the same model from different perspectives. A contemporary April 1993 review describes switching a sales report from sales by product to sales by region or month, and keeping multiple views without writing new formulas. (Los Angeles Times review)

For today’s analyst or data publisher, the useful principle is the same: the source records and the summary shown to readers are different things. A pivot helps answer a question about the records; it should not replace them without explanation.

A small example: records become a summary

Suppose you want to publish spending by region and month. This fictional dataset has one row per spending record, with all amounts in USD:

A pivot with Region in Rows, Month in Columns, and Sum of Amount_USD in Values produces:

Swapping Region and Month changes the arrangement, not the spending records. Microsoft’s documentation explains how fields move between areas and how value fields are summarized. Check that the calculation is Sum, not Count: Excel can default to Count when it interprets the amounts as text. (Microsoft PivotTable guide)

For creation steps, see our Excel pivot-table guide.

Prepare the pivot result for public sharing

Choose what readers need before exporting:

• Publish the records when readers should inspect categories or build different summaries.
• Publish the summary when the page should answer a specific question, such as monthly spending by region.

For a sortable summary dataset, use one row per region-month combination and columns such as Region, Month and Total_spending_USD. Keep grand totals outside those data rows so they cannot be mistaken for additional observations.

If the source has changed, refresh the PivotTable before exporting; Excel’s PivotTables use a cached snapshot of the data. (Microsoft refresh guidance) Then reconcile the summary’s total with the reviewed source, record any filters or exclusions, and explain the measure, currency, period and source. Remove sensitive information; even an aggregate deserves a disclosure review.

TablePage accepts CSV, TSV, XLSX and XLS files and generates a public dataset page with a filterable table. Prepare a clean export of the records or finished summary you intend to share, rather than assuming the workbook’s pivot controls will carry over. Readers should be able to understand what each published row represents without opening a spreadsheet application.