Insurance Excel Data Analysis

Problem

Insurance data in Excel, CSV, or JSON files often needs cleaning, basic statistics, charts, and a written report before it can be shared. This skill turns that routine into a repeatable pipeline: load the file, run exploratory checks, clean obvious issues, generate charts, and export a Word document.

How It Works

It accepts .csv, .xlsx/.xls, and .json, tries common encodings, and reads the first Excel sheet by default. The main workflow is:

• Loading: detect encodings and read the first sheet of multi-sheet workbooks
• Cleaning and statistics: remove blank or duplicate rows, fill missing values, flag outliers, and write summary.json
• Charts: generate bar, pie, line, and horizontal bar charts into charts/*.png
• AI insight: compose a prompt from summary.json, then call Kimi or DeepSeek to generate plain-language insights
• Export: use export_report.py to produce a Word report with data overview, cleaning log, statistics, charts, AI insights, and recommendations

Boundaries

This is best for small-to-medium tabular data and standard reporting, not large-scale warehouse modeling or advanced statistical modeling. AI insight requires an external model API key, and the runtime depends on pandas, matplotlib, numpy, openpyxl, and python-docx. Multi-sheet Excel files default to the first sheet, and complex formulas, pivot tables, or sensitive fields need manual review.