Build a Clean Excel Table for Public Sharing

To turn a spreadsheet into a useful public dataset, start with a rectangular table: one header row, one record per row, and one kind of value per column. Then use Format as Table to make that range an Excel table you can sort, filter and check before sharing.

These are two different things: tabular data describes the arrangement of the records; an Excel table is a workbook feature applied to a cell range. Adding a table style does not repair merged headings, mixed units or duplicate records. Microsoft explains that Excel tables provide sorting, filtering and calculation features for related data. Microsoft’s worksheet guidelines

1. Start with a clear record structure

Enter this fictional sample into cells A1:C5. It represents monthly library visits, with one row per branch per month—not one row per visitor.

Before entering the month labels, format A2:A5 as Text to preserve 2026-01 and 2026-02 as written. Microsoft recommends preformatting cells as Text to prevent automatic date conversion. Date-conversion guidance

month	branch	visits
2026-01	North	40
2026-01	South	35
2026-02	North	45
2026-02	South	30

Keep the publication title and source notes outside the data rectangle. Within it:

• Use one row of descriptive, unique column names.
• Repeat labels such as North on every relevant row; do not merge cells to show groups.
• Keep counts numeric rather than entering strings such as 40 visits.
• Remove blank separator rows and columns.
• Keep subtotals and grand totals outside the records you intend to publish.

Microsoft recommends placing similar items in the same column and avoiding blank rows or columns inside a data range. Data organization guidance

If your starting point is a presentation-style report, use the tabular-data reshaping workflow. If it is a PivotTable, follow the separate Excel tabular-form workflow.

2. Convert the range to an Excel table

In desktop Excel:

1. Select a cell inside the sample data.
2. Choose Home > Format as Table.
3. Pick a style.
4. Check that the proposed range is A1:C5.
5. Select My table has headers, then choose OK.

These are the steps documented in Microsoft’s table-creation guide. For your own dataset, verify that the range includes the last record but excludes titles, notes and totals.

The result is an Excel table around the existing values. The style helps distinguish headers and rows; the structure still needs checking.

3. Check the records, not just the appearance

For this sample, confirm:

• Row count: four records, excluding the header.
• Total: enter =SUM(C2:C5) in E2, outside the table. It should return 150.
• Record uniqueness: each month–branch pair appears once. Repeated branch names alone are expected.
• Filter result: use the branch header’s filter to select only North. Two records should remain visible, with visit counts of 40 and 45. Clear the filter afterward.

For real data, also check missing values and units. A blank visit count should not silently mean zero. Decide whether it means “not reported,” then document that meaning alongside the published table.

Unhide rows and columns and clear filters before reviewing the release copy. This is a review step, not a privacy measure: hiding information is not a substitute for removing it from the file.

4. Prepare a public release file

Create a separate release file containing only the reviewed table—not the check formula in E2. If the table contains calculated results, copy them as values so the release does not depend on the working workbook. Remove sensitive information entirely; do not upload it.

For a single table, CSV is a straightforward export option. In desktop Excel, choose File > Save As and select a CSV file type. Excel saves only the current worksheet and warns that some workbook features are unsupported in text formats. Keep the original XLSX separately. Microsoft’s export instructions

As of October 2026, TablePage accepts CSV, TSV, XLSX and XLS files and generates a public dataset page with a shareable link and filterable table. After uploading the sanitized release, compare the page’s headers, records and values with your checked file.

Supply enough context for readers to interpret it: dataset title, source, period covered, units and missing-value definitions. For this example, identify the figures as fictional sample data—not measured library activity.