Spreadsheets

Nebula Writer lets you keep Excel workbooks and data files in the same workspace as the documents, papers, and slide decks that depend on them.

When to Use This Page#

• You work with .xlsx, .xls, or .csv files
• You want to analyze data beside a report or paper
• You want the assistant to inspect spreadsheets before writing
• You need reviewable AI changes to workbook cells or ranges

Spreadsheet Workspace#

Open a workbook from the project sidebar or drag it into the app.

The spreadsheet view includes:

• Workbook tabs
• Grid navigation
• Formula bar
• Formatting controls
• Sheet switching
• Save/export actions
• Review banners for AI changes

Supported Files#

Keep workbooks in the same project folder as the draft that uses them. The assistant can reason about the relationship between the spreadsheet, the report, and generated figures.

Editing Cells#

Use the grid for direct edits:

1. Click a cell.
2. Type a value or formula.
3. Use the formula bar for longer content.
4. Save the workbook.

Use clear sheet names when the workbook feeds a document. It makes prompts and review easier.

AI Spreadsheet Workflows#

The assistant can inspect workbook structure and propose changes.

Summarize the workbook and identify sheets with formulas.

Use @research-dashboard.xlsx to draft a results summary for @paper.tex.

Find outliers in the Data sheet and suggest one chart for the presentation.

Update the Summary sheet using values from Data, then show me the changed cells.

For multi-sheet workbooks, mention the sheet name when you know it:

Use the Data sheet in @research-dashboard.xlsx. Do not edit the Budget wheel sheet.

Review Spreadsheet Changes#

AI workbook edits should be reviewed before you rely on them.

Look for:

• Changed formulas
• Changed source data
• Hidden rows or columns
• Named range changes
• Sheet-level changes
• Formatting that changes interpretation

If the assistant makes a broad change, reject it and ask for a smaller edit.

Python and Spreadsheets#

Use Python when spreadsheet analysis needs to be reproducible.

import pandas as pd

df = pd.read_excel("office/research-dashboard.xlsx", sheet_name="Data")
summary = df.groupby("Category", as_index=False)["Amount"].sum()
summary.to_csv("exports/category-summary.csv", index=False)

Then ask the assistant to explain or use the output:

Read exports/category-summary.csv and turn it into a concise Results paragraph.

See Python Code for chart and artifact examples.

From Spreadsheet to Writing#

Common workflows:

• Draft a report section from workbook data
• Create a chart and insert it into LaTeX
• Generate a slide summary from a table
• Check whether numbers in a document match the spreadsheet
• Create a cleaned export for collaborators

Prompt example:

Compare @research-dashboard.xlsx with @research-brief.md. Flag any numbers in the brief that do not match the workbook.

Best Practices#

• Keep raw data and derived summaries on separate sheets
• Ask for a plan before changing formulas
• Save before broad AI edits
• Use Python for repeatable analysis
• Review every changed cell
• Export final charts at high resolution for papers and decks