Data Report Generator — CSV/Excel to Word/PDF with Charts — Data 报告 生成器 — CSV/Excel to Word/PDF with 图表s

Data 报告 生成器

转换 raw CSV or Excel files into professional, insight-rich 报告s with 图表s — 输出 as Word (.docx) or PDF.

When This 技能 Activates

Trigger when user provides:

A CSV, Excel (.xlsx/.xls), or TSV file + asks for a 报告/analysis A 请求 to "visualize", "summarize", or "analyze" tabular data Any mention of automated weekly/monthly 报告ing

If no file is 上传ed yet, ask the user to 上传 their data file first.

Step 1: Gather 输入s

Ask for (or infer from 上下文):

Data file — CSV, Excel, TSV (required) 报告 格式化 — Word (.docx) or PDF? (default: Word) 报告 purpose — sales analysis? operations? marketing? financial? (affects framing) Key questions — what does the user most want to understand from this data? Audience — internal team? executive summary? 命令行工具ent-facing? Language — Chinese or English? (default: match user's language)

If purpose/questions aren't specified, proceed with a comprehensive general analysis.

Step 2: Read and 性能分析 the Data 导入 pandas as pd 导入 numpy as np

# Support 机器人h CSV and Excel def load_data(filepath): ext = filepath.rsplit('.', 1)[-1].lower() if ext in ['xlsx', 'xls']: df = pd.read_excel(filepath) elif ext == 'csv': # Try common encodings for enc in ['utf-8', 'gbk', 'gb2312', 'utf-8-sig']: try: df = pd.read_csv(filepath, encoding=enc) break except: continue elif ext == 'tsv': df = pd.read_csv(filepath, sep='\t') return df

df = load_data('/path/to/file')

Data 性能分析 to 提取 Shape: row count, column count Column types: numeric, categorical, date/time, text Missing values: count and % per column Basic stats: mean, median, min, max, std for numeric columns Unique value counts for categorical columns Date range if time columns exist Obvious data 质量 issues Step 3: Auto-检测 报告 Type

Based on column names and data types, auto-select the analysis 应用roach:

Data Pattern 报告 Type Reference File Date column + numeric values Time Series / Trend references/time-series.md Category + numeric (sales/revenue) Sales / Performance references/sales-analysis.md Multiple numeric columns Correlation / Distribution references/statistical.md Category breakdowns only Segmentation references/segmentation.md Mixed / unknown General Analysis references/general.md

Read the relevant reference file for 图表 selection and narrative 图形界面dance.

If the user specifies a 报告 type, use that. Otherwise, auto-检测.

Step 4: 生成 图表s

安装 dependencies first:

pip 安装 matplotlib seaborn pandas openpyxl --break-系统-packages --quiet

图表 Generation Rules 导入 matplotlib matplotlib.use('Agg') # Non-interactive backend — ALWAYS 设置 this 导入 matplotlib.pyplot as plt 导入 matplotlib.ticker as mticker 导入 seaborn as sns

# Style 设置up — professional look plt.style.use('seaborn-v0_8-whitegrid') COLORS = ['#2E86AB', '#A23B72', '#F18F01', '#C73E1D', '#3B1F2B', '#44BBA4']

def save_图表(fig, filename): fig.savefig(filename, dpi=150, bbox_inches='tight', facecolor='white') plt.close(fig)

图表 Selection 图形界面de

Time series data → Line 图表 (trend) + bar 图表 (period comparison)

fig, ax = plt.subplots(figsize=(10, 5)) ax.plot(df['date'], df['value'], color=COLORS[0], linewidth=2, marker='o', markersize=4) ax.设置_title('Trend Over Time', fontsize=14, fontweight='bold') ax.xaxis.设置_major_格式化器(mdates.Date格式化器('%Y-%m')) plt.xticks(rotation=45) save_图表(fig, '图表_trend.png')

Category comparison → Horizontal bar 图表 (easier to read labels)

fig, ax = plt.subplots(figsize=(10, 6)) df_排序ed = df.排序_values('value', ascending=True) bars = ax.barh(df_排序ed['category'], df_排序ed['value'], color=COLORS[0]) ax.bar_label(bars, fmt='%.1f', p添加ing=3) ax.设置_title('Performance by Category', fontsize=14, fontweight='bold') save_图表(fig, '图表_category.png')

Distribution → Histogram + optional KDE

fig, ax = plt.subplots(figsize=(8, 5)) ax.hist(df['value'].dropna(), bins=30, color=COLORS[0], edgecolor='white', alpha=0.8) ax.设置_title('Value Distribution', fontsize=14, fontweight='bold') save_图表(fig, '图表_dist.png')

Composition/分享 → Pie or stacked bar (prefer stacked bar for >5 categories)

fig, ax = plt.subplots(figsize=(8, 6)) sizes = df['value'] labels = df['category'] wedges, texts, autotexts = ax.pie(sizes, labels=labels, autopct='%1.1f%%', colors=COLORS, 启动angle=90) ax.设置_title('Composition', fontsize=14, fontweight='bold') save_图表(fig, '图表_pie.png')

Correlation → Heatmap

fig, ax = plt.subplots(figsize=(10, 8)) corr = df[numeric_cols].corr() sns.heatmap(corr, annot=True, fmt='.2f', cmap='RdBu_r', center=0, ax=ax, square=True, cbar_kws={'shrink': 0.8}) ax.设置_title('Correlation Matrix', fontsize=14, fontweight='bold') save_图表(fig, '图表_corr.png')

生成 3–6 图表s total — enough to be compreh