Messy Table Autopilot
Detects delimiters, encoding, late headers, Excel sheets, empty structure, duplicate fields, and summary rows.
LOCAL-FIRST MESSY TABLE AUTOPILOT
Turn messy tables into explainable analysis.
TablePilot loads messy Excel, CSV, and TXT files, detects structure, infers schema, scores quality, identifies business roles, recommends Chart Studio views, builds a decision brief, exports cleaned data, and generates evidence-grounded reports.
TablePilot 是一个本地优先的复杂表格分析工作台，把混乱表格转化为字段画像、业务角色解读、质量评分、决策简报、动态图表、清洗导出和可解释摘要。
Quality
94 / 100
Autopilot
Clean + Chart
Chart Studio
Auto-selects grouped bars, scatter plots, heatmaps, box plots, and trend views, then ties each chart back to a user-facing decision brief.
Decision Brief
Turns raw profiling into a primary question, segment leaders, driver candidates, evidence, limitations, and suggested next actions.
Bilingual Reports
Chinese mode renders user-facing analysis in Chinese, while English mode stays fully English. Field names remain unchanged as evidence labels.
Clean Export
Compares before/after clean-up, then exports conservative cleaned CSV/XLSX files.
Local AI Optional
Runs deterministically by default, with optional OpenAI-compatible llama.cpp or Ollama wording that never overrides structured evidence.
Product Screenshots
真实桌面端截图：加载销售样例、中文洞察面板，以及清洗前后对比。
Architecture
Qt Desktop
FastAPI
Messy Parser
Business Roles
Clean Export
Decision Brief
Planner
Reports + Charts