categorical-comparison-analysis

This sub-skill covers one capability of the Excel workflow. For reading/counting/Parquet optimization, see the parent workflow SKILL.md.

Step1 读取文件并统计所有 sheet 的总行数，评估是否需要进行大文件优化处理。

import pandas as pd
from pandas import read_excel
from pathlib import Path

# 统计所有 sheet 的行数以决定处理策略
file_path = "input_data.xlsx"
sheet_names = pd.ExcelFile(file_path).sheet_names
total_rows = 0
for sheet in sheet_names:
    # 仅读取行索引以快速计数
    df_tmp = read_excel(file_path, sheet_name=sheet, usecols=[0])
    total_rows += len(df_tmp)

print(f"Total rows across all sheets: {total_rows}")

Step2 提取对比维度的分类信息，执行数据清洗，包括去除空值、处理合并单元格填充以及排除非数据行。

# 定义目标列名
target_col_a = "category_a_column"
target_col_b = "category_b_column"

# 处理合并单元格（ffill）并清洗数据
df[target_col_a] = df[target_col_a].ffill()
df[target_col_b] = df[target_col_b].ffill()

# 排除标题行占位符（如 '代码'、'名称'）及空值
exclude_val = "代码" 
data_a = df[target_col_a].dropna()
data_a = data_a[data_a != exclude_val]

data_b = df[target_col_b].dropna()
data_b = data_b[data_b != exclude_val]

Step3 统计分类数量，计算差异值与占比，生成多维度对比统计表。

count_a = len(data_a)
count_b = len(data_b)
total_count = count_a + count_b
difference = abs(count_a - count_b)

# 计算占比
ratio_a = (count_a / total_count) * 100 if total_count > 0 else 0
ratio_b = (count_b / total_count) * 100 if total_count > 0 else 0

# 构建统计摘要
summary_df = pd.DataFrame({
    "分类名称": ["类别A", "类别B"],
    "数量": [count_a, count_b],
    "占比": [f"{ratio_a:.2f}%", f"{ratio_b:.2f}%"]
})
print(summary_df)
print(f"数量差异: {difference}")

Step4 配置中文字体并生成可视化图表（柱状图与饼图），美化输出效果。

import matplotlib.pyplot as plt

# 中文字体配置
plt.rcParams['font.sans-serif'] = ['SimHei', 'DejaVu Sans']
plt.rcParams['axes.unicode_minus'] = False

fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 6))
labels = ['类别A', '类别B']
counts = [count_a, count_b]
colors = ['#3498db', '#e74c3c']