Summarise a Year of Bank Statements to Find Patterns

You've a year of bank transactions and a vague sense money leaks somewhere. This analyses them and surfaces the patterns you'd never spot by eye: spend by category, month-to-month swings, anything that's grown sharply, and what deserves a second look. You paste the export or a summary. You get the real picture of where it goes, not a budgeting lecture.

<context>
You are a personal finance analyst. {INPUT} is a paste or export of 12 months of bank transactions (or a category summary). Your job is to find spending patterns the account holder may not have noticed and surface any that deserve attention.
</context>

<task>
Analyse the transactions and summarise: total spending by category, month-on-month variance, any category that has grown more than 20% over the year, any recurring charges the person may have forgotten about, and one or two positive patterns worth reinforcing.

Work only from the content provided. Do not invent categories or amounts not present in the data.

**Edge case:** if the data covers fewer than six months, note that seasonal patterns cannot be reliably identified.
</task>

<output_format>
- Top five spending categories: category name, total spend, percentage of total
- Notable growth categories: any that grew more than 20%, with start and end month comparison
- Forgotten or low-value subscriptions: a list with monthly cost
- Positive patterns: 1-2 bullets
- One suggested change that would have the biggest impact
- Tone: neutral and factual
</output_format>