Data analysis in LibreOffice Calc from a sales export

Data analysis in LibreOffice Calc starts with clean transaction rows, then a summary that ties back to the sales export. For a customer-by-month comparison, use real dates, signed net sales and a stable customer identifier. The method below keeps credit notes visible and checks the result before exporting a CSV.

Start with the sales rows you already have

The example is an illustrative US office-supplies distributor reviewing January and February 2026. Use one row per posted invoice or credit note. The invoice ID must identify that document uniquely; if the source contains invoice lines, use a document-and-line key instead.

Put the following fields on a sheet named Sales, with headers in row 1 and records in rows 2 through 9. Column B must contain date values, even though the table displays written dates. Column D contains numbers without embedded currency text.

Net sales exclude sales tax and include the signed credit notes. Keep a separate currency field in a real export and select USD before adding rows. If customer names vary, resolve them to customer IDs before summarizing; an organization and its abbreviated name should not become separate customers.

Import and verify the cell types

Open the CSV through File > Open and review the delimiter, text delimiter and column formats in the import preview. LibreOffice's CSV instructions explain how to set these options and choose a format for each column.

Preserve identifiers as text, especially when they contain leading zeros. Use the date format that matches the export rather than assuming the computer's locale will interpret every date correctly. After import, inspect a known date and a negative amount.

These checks should return 8 for the example's eight records:

=COUNT(Sales.B2:B9)
=COUNT(Sales.D2:D9)

A lower result means at least one expected date or amount is not numeric. That does not identify every possible problem, but it is a useful stop before aggregation. Do not turn missing sales amounts into zero simply to make the check pass.

Build the customer-by-month summary

On a second sheet named Summary, place Customer in A1. Set B1 to =DATE(2026;1;1) and C1 to =DATE(2026;2;1), formatting both as month and year. Put Atlas, Beacon and Cedar in A2:A4.

In B2, enter this formula and fill down through B4, then across to C4:

=SUMIFS(Sales.$D$2:$D$9;Sales.$C$2:$C$9;$A2;Sales.$B$2:$B$9;">="&B$1;Sales.$B$2:$B$9;"<"&DATE(YEAR(B$1);MONTH(B$1)+1;1))

The formula sums net sales only when customer, start date and next-month boundary all match. LibreOffice documents SUMIFS as a sum with multiple criteria. These formulas use Calc's documented semicolon argument separator; use the separator configured in your installation.

The next-month boundary avoids hard-coding the last day of each month. The locked source ranges stay fixed when copied; $A2 keeps the customer column fixed and B$1 keeps the month-header row fixed.

In D2 use =SUM(B2:C2) and fill through D4. In B5 use =SUM(B2:B4) and fill through D5. Atlas's January result is 6,000 - 500 = 5,500; Cedar's February result is 3,000 - 500 = 2,500.

Monthly growth is (14,000 - 12,000) / 12,000 = 16.7%, rounded to one decimal place. In a separate cell, use:

=(C5-B5)/B5

Format it as a percentage. If the starting month's sales are zero, report the dollar movement and explain that a conventional growth percentage is undefined.

Prove that the summaries include the right records

The summary should pass several checks before anyone interprets the growth.

The source-to-summary equality proves that these selected records were allocated to the table. It does not prove the original export is complete: also tie the source total to the sales register or ledger for the same period and scope.

On the Sales sheet, =COUNTIF($A$2:$A$9;A2) filled down should return 1 for each unique document in this example. Investigate repeats before removing anything: repeated invoice IDs can be legitimate when the source has separate line items. For the accounting treatment behind signed transactions, see credit notes and returns in revenue measures.

Use a pivot table when the category list changes often

First label Sales E1 Month. Enter =DATE(YEAR(B2);MONTH(B2);1) in E2 and fill through E9, then select Sales.A1:E9. Choose Insert > Pivot Table, then Current selection. Add customer to Row Fields, Month to Column Fields and net sales to Data Fields. Set the data calculation to Sum; LibreOffice's pivot-table guide explains the layout and calculation settings.

Check the pivot result against the same $26,000 source control total, and refresh it after changing the source. For presenting the checked comparison, use the guide to data analysis graphs and charts.

Where it goes wrong

• Imported text amounts. The table can look correct while sums omit records. Check cell types before accepting the total.
• Positive credit notes. Credits with the wrong sign inflate sales. Preserve the source's documented sign convention.
• Incomplete ranges. New rows outside the formula or pivot source disappear from summaries. Recheck the source boundary after each export.
• Wrong exported sheet. CSV contains the current sheet, not every workbook tab. Choose transaction rows for detailed follow-up or the summary for period-level review.

These are observed sales results. Generalizing them to customers outside the export requires a separate sampling argument, covered in inferential data analysis.

Do this on your own file in Covirage

Save the intended sheet through File > Save As > Text CSV, keep the working workbook separately and reopen the CSV to check dates, signs and totals. Upload that CSV to Covirage: it maps the columns, its tools compute and check the figures, and the external AI model explains the result with every figure traced to its rows. Create a free Covirage account and bring a sample of your own sales export.

See file-based spreadsheet analysis with Covirage for the supporting workflow.

Covirage is not affiliated with LibreOffice or The Document Foundation.