spending-analysis

Spending Analysis

Purpose

Provides rules for analyzing categorized transaction data — aggregating spending by category, comparing month-over-month, detecting unusual transactions, and generating savings recommendations.

Category Aggregation

For a given month, aggregate all categorized transactions:

1. Group transactions by category
2. For each category, compute:
   • total: Sum of all transaction amounts (absolute values for expenses)
   • transaction_count: Number of transactions
   • avg_transaction: total / transaction_count
3. Calculate each category's percentage of total spending: category_total / total_spending × 100
4. Sort categories by total (highest first)

Write results to monthly-summary.csv.

Receipt-Level Breakdown

For a given merchant, aggregate spending at the receipt-item level instead of the transaction level. Used when the user asks about spending at a specific merchant (e.g., "Hvad bruger jeg hos Bilka?").

Data Source

1. Read receipts.csv, filter by merchant (case-insensitive match against normalized merchant name per skills/categorization/SKILL.md)
2. If month argument is provided, also filter by date column (YYYY-MM prefix match)
3. Collect all matching receipt_id values and their date values
4. Determine which monthly files to read: extract unique YYYY-MM values from the receipt dates
5. For each monthly file receipt-items-{YYYY-MM}.csv: read it if it exists, filter to rows where receipt_id is in the collected set. Skip missing files.
6. Combine all matching rows

Aggregation Rules

Group matching receipt items by subcategory:

1. For each subcategory, compute:
   • total: Sum of total_price for all items in that subcategory
   • item_count: Sum of quantity for all items in that subcategory
   • avg_price: total / item_count
2. Calculate each subcategory's percentage of total merchant spending: subcategory_total / merchant_total × 100
3. Sort subcategories by total (highest first)