watch-my-money

# watch-my-money

Analyze transactions, categorize spending, track budgets, flag overspending.

## Workflow

### 1. Get Transactions

Ask user for bank/card CSV export OR pasted text.

Common sources:
- Download CSV from your bank's online portal
- Export from budgeting apps
- Copy/paste transactions from statements

Supported formats:
- Any CSV with date, description, amount columns
- Pasted text: "2026-01-03 Starbucks -5.40 CHF"

### 2. Parse & Normalize

Read input, normalize to standard format:
- Auto-detect delimiter (comma, semicolon, tab)
- Parse dates (YYYY-MM-DD, DD/MM/YYYY, MM/DD/YYYY)
- Normalize amounts (expenses negative, income positive)
- Extract merchant from description
- Detect recurring transactions (subscriptions)

### 3. Categorize Transactions

For each transaction, assign category:

**Categories:**
- rent, utilities, subscriptions, groceries, eating_out
- transport, travel, shopping, health
- income, transfers, other

Categorization order:
1. Check saved merchant overrides
2. Apply deterministic keyword rules (see [common-merchants.md](references/common-merchants.md))
3. Pattern matching (subscriptions, utilities)
4. Heuristic fallback

For ambiguous merchants (batch of 5-10), ask user to confirm.
Save overrides for future runs.

### 4. Check Budgets

Compare spending against user-defined budgets.

Alert thresholds:
- 80% - approaching limit (yellow)
- 100% - at limit (red)
- 120% - over budget (red, urgent)

See [budget-templates.md](references/budget-templates.md) for suggested budgets.

### 5. Detect Anomalies

Flag unusual spending:
- Category spike: spend > 1.5x baseline AND delta > 50
- Subscription growth: subscriptions up > 20%
- New expensive merchant: first appearance AND spend > 30
- Potential subscriptions: recurring same-amount charges

Baseline = previous 3 months average (or current month if no history).

### 6. Generate HTML Report

Create local HTML file with:
- Month summary (income, expenses, net)
- Category breakdown with b...