Unified Customer Service Data Dashboard

Problem context

Customer service data is scattered across Google Sheets, Tencent Docs, Zendesk/Intercom, WeCom, MySQL/PostgreSQL, Excel/CSV, and inconsistent field names make metrics drift. Teams need a reliable way to produce shareable, report-ready, and collaborative dashboards without manually assembling tables each time.

How it works

• Multi-source intake: read data via read_google_sheet.py, pandas.read_sql, APIs, or tencent-docs MCP, then normalize it into a standard DataFrame.
• Metric calculation: covers inquiry volume, ticket volume, average response time, SLA compliance, CSAT, NPS, FCR, escalation rate, per-agent capacity, and more; user-selected metrics take priority, otherwise outputs the full set.
• Dashboard output: generate an interactive HTML dashboard, Excel workbook, PPT deck, or Tencent Docs smart sheet for browser sharing, archiving, meetings, and team collaboration.
• Scheduled refresh: configure daily/weekly regeneration; for Google Sheets, enable --live-sheet-id so the HTML page fetches the latest data when opened.

Boundaries

Requires stable data sources and clear field semantics. Missing fields such as ticket_id, created_at, or csat_score cause some metrics to show N/A. For datasets over 100k rows, sample or filter by date first. For performance data, sensitive fields, API tokens, or Sheet sharing permissions, confirm data safety and access validity before publishing.