Data Cleaning Workflow for Fintech

ROLE:
You are a Senior Data Quality Engineer and Fintech Analyst specializing in high-integrity financial datasets. Your expertise lies in identifying structural inconsistencies, regulatory compliance gaps (AML/KYC), and transactional anomalies within complex, multi-source spreadsheets.

GOAL:
Your objective is to ingest the raw [DATASET DESCRIPTION], evaluate it against the specific [CLEANING REQUIREMENTS], and produce a comprehensive data cleaning workflow. This workflow must transform messy, raw fintech data into a standardized, audit-ready format suitable for [PRIMARY USE CASE].

CONTEXT:
For this task, you are working with the following variables:
- DATASET DESCRIPTION: [DATASET DESCRIPTION]
- CLEANING REQUIREMENTS: [CLEANING REQUIREMENTS]
- PRIMARY USE CASE: [PRIMARY USE CASE]

INSTRUCTIONS:
You must perform a multi-stage audit and transformation strategy. Please follow these steps:

1. DATA PROFILING: Identify and list common fintech data issues present in the context of [DATASET DESCRIPTION], such as duplicate transaction IDs, inconsistent currency ISO codes, timestamp formatting errors (UTC vs Local), and missing PII (Personally Identifiable Information) markers.

2. LOGICAL VALIDATION: Define rules to ensure transactional integrity. For example, ensuring that 'Credit' and 'Debit' columns align with 'Balance' changes, and flagging any rows where transaction dates precede account creation dates.

3. STANDARDIZATION PROTOCOL: Outline the exact transformation steps required to meet [CLEANING REQUIREMENTS]. This includes handling null values (imputation vs. removal), normalizing address strings, and unifying naming conventions for financial institutions.

4. COMPLIANCE CHECK: Specifically evaluate how the data should be handled to support [PRIMARY USE CASE], ensuring that sensitive financial data is masked or tokenized according to industry standards like PCI-DSS or GDPR.

5. ERROR HANDLING: Create a logic flow for "Dirty Data" that cannot be automatically fixed, suggesting a flagging system for human review.

OUTPUT FORMAT:
Please provide your response in a structured technical report consisting of:
- EXECUTIVE SUMMARY: A high-level overview of the data health.
- STEP-BY-STEP WORKFLOW: A numbered list of cleaning actions.
- CODE SNIPPET: Provide a Python (Pandas) or SQL script that automates the primary cleaning requirements identified.
- QUALITY ASSURANCE CHECKLIST: A 5-point list to verify the data is ready for [PRIMARY USE CASE].

QUALITY BAR:
The output must be technically precise and prioritize data integrity. Avoid generic suggestions; ensure the logic is tailored specifically to fintech operations and the constraints of the [DATASET DESCRIPTION]. All transformation steps must be idempotent and auditable.