Complaint Quality Review Analysis

Problem

Enterprise complaint ledgers often sit across worksheets with inconsistent field names, making manual review slow and prone to missed repeat issues. This skill targets after-sales quality review, turning Excel, CSV, or pasted complaint records into structured analysis. It first creates a macro data overview, then confirms key fields such as issue level and resolution status, reducing field-mapping errors. It fits annual reviews, single-customer complaint concentration analysis, and monthly top-issue reporting, but does not handle shop-floor defects, incoming supplier materials, SPC, or cost/process analysis.

How It Works

The workflow is confirm first, compute second, generate last. The skill parses all worksheets and identifies fields, types, and row counts. It then shows preliminary insights, including date range, complaint volume, major customers and product lines, plus data-quality flags such as missing fields or invalid dates. Next, it runs category statistics, customer/product distribution, and repeated-complaint detection, marking issues when the same keyword, customer, and time window appear at least three times. It finally outputs report.html and report.md; the HTML report embeds matplotlib charts, while Markdown is easy to paste into weekly reports or review documents. The report keeps a fixed structure: period overview, category statistics and high-frequency issues, and quality conclusions with internal-control recommendations.