XLSX — Spreadsheet Skill

Two approaches for working with spreadsheets depending on the task:

Approach	Tool	Best For
Programmatic	openpyxl / pandas	Creating files, bulk data, formulas, automation
Browser-based	Excel Online via Playwright	Pivot tables, charts, collaboration, interactive editing

---

Part 1: Programmatic Approach (openpyxl / pandas)

Requirements for All Excel Outputs

Professional Font

• Use a consistent, professional font (e.g., Arial, Times New Roman) unless otherwise instructed

Zero Formula Errors

• Every Excel model MUST be delivered with ZERO formula errors (#REF!, #DIV/0!, #VALUE!, #N/A, #NAME?)

Preserve Existing Templates

• Study and EXACTLY match existing format, style, and conventions when modifying files
• Existing template conventions ALWAYS override these guidelines

Financial Model Standards

Color Coding (Industry-Standard)

• Blue text (RGB: 0,0,255): Hardcoded inputs and scenario-adjustable numbers
• Black text (RGB: 0,0,0): ALL formulas and calculations
• Green text (RGB: 0,128,0): Links pulling from other worksheets
• Red text (RGB: 255,0,0): External links to other files
• Yellow background (RGB: 255,255,0): Key assumptions needing attention

Number Formatting

• Years: Format as text strings ("2024" not "2,024")
• Currency: Use $#,##0 format; ALWAYS specify units in headers ("Revenue ($mm)")
• Zeros: Use formatting to display as "-" including percentages
• Percentages: Default to 0.0% format (one decimal)
• Multiples: Format as 0.0x for valuation multiples
• Negative numbers: Use parentheses (123) not minus -123

Formula Rules

• Place ALL assumptions in separate cells, use cell references in formulas
• Use =B5*(1+$B$6) instead of =B5*1.05
• Document hardcodes with source comments

CRITICAL: Use Formulas, Not Hardcoded Values

Always use Excel formulas instead of calculating values in Python and hardcoding them.

# WRONG - Hardcoding calculated values
total = df['Sales'].sum()
sheet['B10'] = total  # Hardcodes 5000

# CORRECT - Using Excel formulas
sheet['B10'] = '=SUM(B2:B9)'
sheet['C5'] = '=(C4-C2)/C2'
sheet['D20'] = '=AVERAGE(D2:D19)'

Reading and Analyzing Data

import pandas as pd

# Read Excel
df = pd.read_excel('file.xlsx')                    # Default: first sheet
all_sheets = pd.read_excel('file.xlsx', sheet_name=None)  # All sheets as dict

# Analyze
df.head()      # Preview data
df.info()      # Column info
df.describe()  # Statistics

# Write Excel
df.to_excel('output.xlsx', index=False)

Creating New Excel Files

from openpyxl import Workbook
from openpyxl.styles import Font, PatternFill, Alignment

wb = Workbook()
sheet = wb.active

# Add data
sheet['A1'] = 'Hello'
sheet['B1'] = 'World'
sheet.append(['Row', 'of', 'data'])

# Add formula
sheet['B2'] = '=SUM(A1:A10)'

# Formatting
sheet['A1'].font = Font(bold=True, color='FF0000')
sheet['A1'].fill = PatternFill('solid', start_color='FFFF00')
sheet['A1'].alignment = Alignment(horizontal='center')

# Column width
sheet.column_dimensions['A'].width = 20

wb.save('output.xlsx')

Editing Existing Files

from openpyxl import load_workbook

wb = load_workbook('existing.xlsx')
sheet = wb.active  # or wb['SheetName']

# Working with multiple sheets
for sheet_name in wb.sheetnames:
    sheet = wb[sheet_name]

# Modify cells
sheet['A1'] = 'New Value'
sheet.insert_rows(2)
sheet.delete_cols(3)

# Add new sheet
new_sheet = wb.create_sheet('NewSheet')
new_sheet['A1'] = 'Data'

wb.save('modified.xlsx')

Recalculating Formulas

LibreOffice Required: Use the provided scripts/recalc.py script to recalculate formula values.

python scripts/recalc.py output.xlsx 30

The script:

• Automatically sets up LibreOffice macro on first run
• Recalculates all formulas in all sheets
• Scans ALL cells for Excel errors
• Returns JSON with detailed error locations and counts

{
  "status": "success",
  "total_errors": 0,
  "total_formulas": 42,
  "error_summary": {}
}

Formula Verification Checklist

• Test 2-3 sample references before building full model
• Confirm column mapping (column 64 = BL, not BK)
• Remember row offset (DataFrame row 5 = Excel row 6)
• Handle NaN with pd.notna()
• Check far-right columns (FY data often in columns 50+)
• Verify cross-sheet references use correct format (Sheet1!A1)
• Test edge cases: zero, negative, and very large values

Library Selection Guide

• pandas: Best for data analysis, bulk operations, simple data export
• openpyxl: Best for complex formatting, formulas, and Excel-specific features
• Warning: load_workbook('file.xlsx', data_only=True) replaces formulas with values if saved

---

Part 2: Browser-Based Approach (Excel Online)

For interactive spreadsheet work via Playwright MCP automation.

Setup

Configure via canifi-env:

canifi-env set MICROSOFT_EMAIL "your-email@outlook.com"

Authentication Options

Option 1: Manual Browser Login (Recommended)

1. Complete Browser Automation Setup using CDP mode
2. Login to Excel Online manually in the Playwright-controlled Chrome window
3. Claude uses your authenticated session without seeing your password

Option 2: Environment Variables

canifi-env set SERVICE_EMAIL "your-email"
canifi-env set SERVICE_PASSWORD "your-password"

Browser Capabilities

• Create and edit spreadsheets
• Build formulas and functions
• Create charts and visualizations
• Apply conditional formatting
• Create pivot tables
• Sort and filter data
• Use data validation
• Import and export data
• Collaborate in real-time

Usage Examples

Create Spreadsheet:

User: "Create an expense tracker in Excel"
Claude: Creates new workbook "Expense Tracker", adds columns for
        Date, Category, Description, Amount. Adds SUM formula for total.

Analyze Data:

User: "Create a pivot table from my sales data"
Claude: Selects data range, inserts pivot table,
        configures rows, columns, and values.

Create Chart:

User: "Make a line chart showing revenue trends"
Claude: Selects revenue data, inserts line chart,
        adds titles and labels.

Selectors Reference

// New workbook
'[aria-label="New blank workbook"]'

// Cell input
'.formulabar-input' or 'input[name="Cell"]'

// Ribbon tabs
'[role="tablist"]'
'[aria-label="Insert"]'
'[aria-label="Formulas"]'
'[aria-label="Data"]'

// Actions
'[aria-label="Insert chart"]'
'[aria-label="Sort"]'
'[aria-label="Filter"]'

Common Formulas

=SUM(A1:A10)              // Sum range
=AVERAGE(A1:A10)          // Average
=VLOOKUP(key,range,col,0) // Vertical lookup
=IF(condition,true,false)  // Conditional
=COUNTIF(range,criteria)   // Count matching
=SUMIF(range,crit,sum)     // Sum matching
=TEXT(A1,"format")         // Format text
=TODAY()                   // Current date
=CONCATENATE(A1,B1)        // Join text

Error Handling

• Login Failed: Retry 3 times, notify user
• Session Expired: Re-authenticate automatically
• Workbook Not Found: Search OneDrive, ask for clarification
• Formula Error: Identify error type, suggest fix
• Save Failed: Enable AutoSave, retry

Excel Online Limitations

• Auto-saves to OneDrive
• Some advanced features only in desktop version
• Maximum rows: 1,048,576 per sheet
• Power Query limited in online version
• Macros not supported in online version
• Can open and edit .xlsx, .xlsm files

---

Code Style Guidelines

For Python code: Write minimal, concise code without unnecessary comments or verbose variable names.

For Excel files: Add comments to cells with complex formulas, document data sources for hardcoded values, include notes for key calculations.