changan-su/forsion/data-analysis-python

import pandas as pd
df = pd.read_csv(path)  # 编码报错时依次试 encoding='utf-8-sig' / 'gbk'
print(df.shape)
print(df.dtypes)
print(df.head(10))
print(df.isna().sum())
print(df.describe(include='all').T)

# 列名规整
df.columns = df.columns.str.strip()
# 类型转换:失败置 NaN 而不是抛错,便于统计坏行
df['amount'] = pd.to_numeric(df['amount'], errors='coerce')
df['date'] = pd.to_datetime(df['date'], errors='coerce')
bad = df['amount'].isna().sum()
print(f'无法解析的 amount 行数: {bad}')   # 坏数据要报告,不能静默丢弃
# 去重与缺失处理(策略要在结论里说明)
df = df.drop_duplicates()
df = df.dropna(subset=['date'])           # 关键列缺失才删行

# 分组聚合
g = (df.groupby('category')
       .agg(总额=('amount','sum'), 单数=('amount','count'), 均值=('amount','mean'))
       .sort_values('总额', ascending=False))
print(g.round(2))
# 时间序列:按月汇总
m = df.set_index('date').resample('MS')['amount'].sum()
print(m)
# 同比/环比
print(m.pct_change().round(4))
# 透视表
p = pd.pivot_table(df, index='region', columns='category',
                   values='amount', aggfunc='sum', fill_value=0)

import matplotlib
import matplotlib.pyplot as plt
from matplotlib import font_manager
# 按优先级找系统里存在的中文字体
for f in ['Noto Sans CJK SC','WenQuanYi Micro Hei','SimHei','PingFang SC','Microsoft YaHei']:
    if any(f in x.name for x in font_manager.fontManager.ttflist):
        matplotlib.rcParams['font.sans-serif'] = [f]
        break
matplotlib.rcParams['axes.unicode_minus'] = False  # 负号正常显示

fig, ax = plt.subplots(figsize=(10, 6), dpi=120)
g['总额'].plot(kind='bar', ax=ax)
ax.set_title('各品类销售总额')
ax.set_xlabel('品类'); ax.set_ylabel('金额(元)')
plt.tight_layout()
out = '/tmp/category_sales.png'
plt.savefig(out); plt.close()
print(f'图表已保存: {out}')

agg = {}
for chunk in pd.read_csv(path, chunksize=200_000, usecols=['category','amount']):
    s = chunk.groupby('category')['amount'].sum()
    for k, v in s.items():
        agg[k] = agg.get(k, 0) + v
result = pd.Series(agg).sort_values(ascending=False)