130 lines
2.7 KiB
Markdown
130 lines
2.7 KiB
Markdown
# Excel 数据分析
|
||
|
||
> ⚠️ **使用本文档前请注意**：本文档应在实际分析 Excel 数据之前阅读，以了解正确的 pandas 分析方法。请先阅读 excel_reading.md 学习如何读取数据。
|
||
|
||
使用 pandas 对 Excel 数据进行常规分析操作。
|
||
|
||
## 快速参考
|
||
|
||
| 任务 | 常用方法 | 代码示例 |
|
||
|------|----------|----------|
|
||
| 按条件过滤 | 布尔索引 | `df[df['sales'] > 10000]` |
|
||
| 分组聚合 | groupby | `df.groupby('region')['sales'].sum()` |
|
||
| 排序 | sort_values | `df.sort_values('sales', ascending=False)` |
|
||
| 计算新列 | 直接赋值 | `df['profit'] = df['revenue'] - df['cost']` |
|
||
| 统计汇总 | describe | `df.describe()` |
|
||
|
||
## 分组聚合（GroupBy）
|
||
|
||
```python
|
||
import pandas as pd
|
||
|
||
df = pd.read_excel("sales.xlsx")
|
||
|
||
# 按列分组并聚合
|
||
sales_by_region = df.groupby("region")["sales"].sum()
|
||
print(sales_by_region)
|
||
|
||
# 多列分组和多重聚合
|
||
result = df.groupby(["region", "product"]).agg({
|
||
"sales": "sum",
|
||
"quantity": "count",
|
||
"price": "mean"
|
||
})
|
||
```
|
||
|
||
## 数据过滤
|
||
|
||
```python
|
||
# 按条件过滤行
|
||
high_sales = df[df["sales"] > 10000]
|
||
|
||
# 多条件过滤
|
||
filtered = df[(df["sales"] > 10000) & (df["region"] == "North")]
|
||
|
||
# 使用 isin 过滤
|
||
selected = df[df["product"].isin(["A", "B", "C"])]
|
||
```
|
||
|
||
## 派生指标计算
|
||
|
||
```python
|
||
# 计算新列
|
||
df["profit_margin"] = (df["revenue"] - df["cost"]) / df["revenue"]
|
||
|
||
# 百分比计算
|
||
df["growth_rate"] = (df["current"] - df["previous"]) / df["previous"] * 100
|
||
|
||
# 累计求和
|
||
df["cumulative_sales"] = df["sales"].cumsum()
|
||
```
|
||
|
||
## 排序
|
||
|
||
```python
|
||
# 按单列排序
|
||
df_sorted = df.sort_values("sales", ascending=False)
|
||
|
||
# 按多列排序
|
||
df_sorted = df.sort_values(["region", "sales"], ascending=[True, False])
|
||
```
|
||
|
||
## 数据透视表
|
||
|
||
```python
|
||
# 创建数据透视表
|
||
pivot = pd.pivot_table(
|
||
df,
|
||
values="sales",
|
||
index="region",
|
||
columns="product",
|
||
aggfunc="sum",
|
||
fill_value=0
|
||
)
|
||
|
||
print(pivot)
|
||
```
|
||
|
||
## 统计分析
|
||
|
||
```python
|
||
# 基本统计
|
||
print(df.describe())
|
||
|
||
# 特定列统计
|
||
print(df["sales"].mean())
|
||
print(df["sales"].median())
|
||
print(df["sales"].std())
|
||
|
||
# 计数统计
|
||
print(df["category"].value_counts())
|
||
```
|
||
|
||
## 数据合并
|
||
|
||
```python
|
||
# 垂直合并多个 DataFrame
|
||
combined = pd.concat([df1, df2], ignore_index=True)
|
||
|
||
# 按公共列合并（类似 SQL JOIN）
|
||
merged = pd.merge(sales, customers, on="customer_id", how="left")
|
||
```
|
||
|
||
## 数据清洗
|
||
|
||
```python
|
||
# 删除重复行
|
||
df = df.drop_duplicates()
|
||
|
||
# 处理缺失值
|
||
df = df.fillna(0) # 填充为 0
|
||
df = df.dropna() # 删除含缺失值的行
|
||
|
||
# 去除空格
|
||
df["name"] = df["name"].str.strip()
|
||
|
||
# 类型转换
|
||
df["date"] = pd.to_datetime(df["date"])
|
||
df["amount"] = pd.to_numeric(df["amount"], errors="coerce")
|
||
```
|