这里写自定义目录标题
- Pandas 示例
- NumPy 示例
- 数据分析案例: Iris 数据集
Pandas 示例
Pandas 是一个数据处理和分析的库，它提供了 DataFrame 和 Series 这两种主要数据结构。
# 从字典创建 DataFrame
import pandas as pd# 创建一个字典
data = {'Name': ['Alice', 'Bob', 'Charlie', 'a'],'Age': [25, 30, 35, None],'City': ['New York', 'Paris', 'London', 'London']
}# 使用字典创建 DataFrame
df = pd.DataFrame(data)
df.to_csv('dict.csv', index=False)
# 显示 DataFrame
print(df) Name Age City
0 Alice 25.0 New York
1 Bob 30.0 Paris
2 Charlie 35.0 London
3 a NaN London
# 读取 CSV 文件
df = pd.read_csv('dict.csv')# 显示前五行
print(df.head()) Name Age City
0 Alice 25.0 New York
1 Bob 30.0 Paris
2 Charlie 35.0 London
3 a NaN London
# 数据过滤
# 筛选年龄大于 30 的行
filtered_df = df[df['Age'] > 30]# 显示结果
print(filtered_df) Name Age City
2 Charlie 35.0 London
# 数据排序
df_sorted = df.sort_values(by='Age', ascending=False)
print(df_sorted) Name Age City
2 Charlie 35.0 London
1 Bob 30.0 Paris
0 Alice 25.0 New York
3 a NaN London
# 缺失值处理
df['Age'] = df['Age'].fillna(df['Age'].mean())
print(df) Name Age City
0 Alice 25.0 New York
1 Bob 30.0 Paris
2 Charlie 35.0 London
3 a 30.0 London
# 数据分组和聚合
grouped = df.groupby('City').count()
print(grouped) Name Age
City
London 2 2
New York 1 1
Paris 1 1
# 数据连接（合并）
other_data = pd.DataFrame({'Name': ['Alice', 'Bob'], 'Salary': [70000, 80000]})
merged_df = pd.merge(df, other_data, on='Name')
print(merged_df) Name Age City Salary
0 Alice 25.0 New York 70000
1 Bob 30.0 Paris 80000
# 应用函数
df['Age'] = df['Age'].apply(lambda x: x + 1)
print(df) Name Age City
0 Alice 26.0 New York
1 Bob 31.0 Paris
2 Charlie 36.0 London
3 a 31.0 London
# 时间序列处理
times = pd.date_range('2020-01-01', periods=3, freq='D')
time_df = pd.DataFrame({'Date': times, 'Value': [1, 2, 3]})
print(time_df) Date Value
0 2020-01-01 1
1 2020-01-02 2
2 2020-01-03 3
# 数据透视表
pivot = df.pivot_table(values='Age', index='City', aggfunc='mean')
print(pivot) Age
City
London 33.5
New York 26.0
Paris 31.0
# 数据可视化（使用 Pandas 内置功能）
import matplotlib.pyplot as pltdf.plot(kind='bar', x='Name', y='Age')
plt.show()
# 读取 JSON 数据
json_data = '''
[{"Name": "Alice", "Age": 25, "City": "New York"},{"Name": "Bob", "Age": 30, "City": "Los Angeles"}
]
'''
with open('data.json', 'w') as file:file.write(json_data)df_json = pd.read_json('data.json')
print(df_json) Name Age City
0 Alice 25 New York
1 Bob 30 Los Angeles
#数据导出到 CSV
df.to_csv('exported_data.csv', index=False)# 多条件过滤
df_filtered = df[(df['Age'] > 25) & (df['City'] == 'New York')]
print(df_filtered) Name Age City
0 Alice 26.0 New York
# 数据列转换
df['AgeSquared'] = df['Age'] ** 2
print(df) Name Age City AgeSquared
0 Alice 26.0 New York 676.0
1 Bob 31.0 Paris 961.0
2 Charlie 36.0 London 1296.0
3 a 31.0 London 961.0
# 唯一值和值计数
print(df['City'].unique())
print(df['City'].value_counts())['New York' 'Paris' 'London']
City
London 2
New York 1
Paris 1
Name: count, dtype: int64
# 缺失值检查
print(df.isnull().sum())Name 0
Age 0
City 0
AgeSquared 0
dtype: int64
# 行列变换
df_transposed = df.T
print(df_transposed) 0 1 2 3
Name Alice Bob Charlie a
Age 26.0 31.0 36.0 31.0
City New York Paris London London
AgeSquared 676.0 961.0 1296.0 961.0
# 字符串操作
df['Name'] = df['Name'].str.upper()
print(df) Name Age City AgeSquared
0 ALICE 26.0 New York 676.0
1 BOB 31.0 Paris 961.0
2 CHARLIE 36.0 London 1296.0
3 A 31.0 London 961.0
# 分类数据处理
df['Category'] = pd.Categorical(df['City'])
print(df) Name Age City AgeSquared Category
0 ALICE 26.0 New York 676.0 New York
1 BOB 31.0 Paris 961.0 Paris
2 CHARLIE 36.0 London 1296.0 London
3 A 31.0 London 961.0 London
# 数据框索引操作
df.set_index('Name', inplace=True)
print(df) Age City AgeSquared Category
Name
ALICE 26.0 New York 676.0 New York
BOB 31.0 Paris 961.0 Paris
CHARLIE 36.0 London 1296.0 London
A 31.0 London 961.0 London
# 多索引数据处理
# 创建一个带有多重索引的 DataFrame
multi_index_df = pd.DataFrame({'A': range(4),'B': range(4, 8)
})
multi_index_df = multi_index_df.set_index([pd.Index(['one', 'two', 'three', 'four']), 'A'])print(multi_index_df) BA
one 0 4
two 1 5
three 2 6
four 3 7
# 时间序列数据重采样
# 创建一个时间序列数据
ts_df = pd.DataFrame({'Date': pd.date_range(start='2021-01-01', periods=4, freq='M'),'Value': [10, 20, 30, 40]
})
ts_df = ts_df.set_index('Date')# 月数据重采样为季度数据，取平均值
resampled_df = ts_df.resample('Q').mean()print(resampled_df) Value
Date
2021-03-31 20.0
2021-06-30 40.0
# 数据窗口函数
# 创建数据
import numpy as npwindow_df = pd.DataFrame({'B': [0, 1, 2, np.nan, 4]})# 使用窗口函数计算滚动平均值
window_df['rolling_mean'] = window_df['B'].rolling(window=2).mean()print(window_df) B rolling_mean
0 0.0 NaN
1 1.0 0.5
2 2.0 1.5
3 NaN NaN
4 4.0 NaN
# 条件式数据替换
# 使用 np.where 来进行条件替换
df['New_Column'] = np.where(df['Age'] > 30, 'Over 30', '30 or under')print(df) Age City AgeSquared Category New_Column
Name
ALICE 26.0 New York 676.0 New York 30 or under
BOB 31.0 Paris 961.0 Paris Over 30
CHARLIE 36.0 London 1296.0 London Over 30
A 31.0 London 961.0 London Over 30
# 数据分组转换
# 使用 transform() 方法
df['Age_Mean_By_City'] = df.groupby('City')['Age'].transform('mean')print(df) Age City AgeSquared Category New_Column Age_Mean_By_City
Name
ALICE 26.0 New York 676.0 New York 30 or under 26.0
BOB 31.0 Paris 961.0 Paris Over 30 31.0
CHARLIE 36.0 London 1296.0 London Over 30 33.5
A 31.0 London 961.0 London Over 30 33.5
# 时间序列数据的移动
# 移动（shift）时间序列数据
ts_df['Previous_Value'] = ts_df['Value'].shift(1)print(ts_df) Value Previous_Value
Date
2021-01-31 10 NaN
2021-02-28 20 10.0
2021-03-31 30 20.0
2021-04-30 40 30.0
# 数据排名
# 使用 rank() 方法
df['Age_Rank'] = df['Age'].rank(method='dense')print(df)
df.columns Age City AgeSquared Category New_Column Age_Mean_By_City \
Name
ALICE 26.0 New York 676.0 New York 30 or under 26.0
BOB 31.0 Paris 961.0 Paris Over 30 31.0
CHARLIE 36.0 London 1296.0 London Over 30 33.5
A 31.0 London 961.0 London Over 30 33.5 Age_Rank
Name
ALICE 1.0
BOB 2.0
CHARLIE 3.0
A 2.0 Index(['Age', 'City', 'AgeSquared', 'Category', 'New_Column','Age_Mean_By_City', 'Age_Rank'],dtype='object')
# # 复杂的字符串操作
# 复杂的字符串操作
df['Name_Length'] = df['City'].str.len()print(df) Age City AgeSquared Category New_Column Age_Mean_By_City \
Name
ALICE 26.0 New York 676.0 New York 30 or under 26.0
BOB 31.0 Paris 961.0 Paris Over 30 31.0
CHARLIE 36.0 London 1296.0 London Over 30 33.5
A 31.0 London 961.0 London Over 30 33.5 Age_Rank Name_Length
Name
ALICE 1.0 8
BOB 2.0 5
CHARLIE 3.0 6
A 2.0 6
# 数据聚合
# 使用更复杂的聚合函数
agg_df = df.groupby('City').agg({'Age': ['mean', 'min', 'max']})agg_df| Age | |||
|---|---|---|---|
| mean | min | max | |
| City | |||
| London | 33.5 | 31.0 | 36.0 |
| New York | 26.0 | 26.0 | 26.0 |
| Paris | 31.0 | 31.0 | 31.0 |
# 合并数据时的不同连接类型
# 创建第二个数据集
other_df = pd.DataFrame({'Name': ['Alice', 'Charlie'], 'Income': [60000, 80000]})# 执行外连接合并
merged_df = pd.merge(df, other_df, on='Name', how='outer')merged_df| Name | Age | City | AgeSquared | Category | New_Column | Age_Mean_By_City | Age_Rank | Name_Length | Income | |
|---|---|---|---|---|---|---|---|---|---|---|
| 0 | ALICE | 26.0 | New York | 676.0 | New York | 30 or under | 26.0 | 1.0 | 8.0 | NaN |
| 1 | BOB | 31.0 | Paris | 961.0 | Paris | Over 30 | 31.0 | 2.0 | 5.0 | NaN |
| 2 | CHARLIE | 36.0 | London | 1296.0 | London | Over 30 | 33.5 | 3.0 | 6.0 | NaN |
| 3 | A | 31.0 | London | 961.0 | London | Over 30 | 33.5 | 2.0 | 6.0 | NaN |
| 4 | Alice | NaN | NaN | NaN | NaN | NaN | NaN | NaN | NaN | 60000.0 |
| 5 | Charlie | NaN | NaN | NaN | NaN | NaN | NaN | NaN | NaN | 80000.0 |
# 多条件数据查询
# 多条件查询
df_query = df.query('Age > 25 and City == "New York"')
print(df_query) Age City AgeSquared Category New_Column Age_Mean_By_City \
Name
ALICE 26.0 New York 676.0 New York 30 or under 26.0 Age_Rank Name_Length
Name
ALICE 1.0 8
# 数据列的分割
# 假设存在一个合并的列，例如 'Name_Age', 需要将其分割为两个独立的列
df['Name_Age'] = df.index + '_' + df['Age'].astype(str)
df[['Split_Name', 'Split_Age']] = df['Name_Age'].str.split('_', expand=True)
print(df) Age City AgeSquared Category New_Column Age_Mean_By_City \
Name
ALICE 26.0 New York 676.0 New York 30 or under 26.0
BOB 31.0 Paris 961.0 Paris Over 30 31.0
CHARLIE 36.0 London 1296.0 London Over 30 33.5
A 31.0 London 961.0 London Over 30 33.5 Age_Rank Name_Length Name_Age Split_Name Split_Age
Name
ALICE 1.0 8 ALICE_26.0 ALICE 26.0
BOB 2.0 5 BOB_31.0 BOB 31.0
CHARLIE 3.0 6 CHARLIE_36.0 CHARLIE 36.0
A 2.0 6 A_31.0 A 31.0
# 数据透视表的高级应用
# 更复杂的数据透视表
pivot_table = df.pivot_table(values='Age', index='City', columns='Name', aggfunc='mean', fill_value=0)
print(pivot_table)Name A ALICE BOB CHARLIE
City
London 31.0 0.0 0.0 36.0
New York 0.0 26.0 0.0 0.0
Paris 0.0 0.0 31.0 0.0
# 使用 at 和 iat 访问单个元素
# 使用 at 和 iat 访问和修改数据框中的特定元素
df.at[0, 'Age'] = 26 # 使用行标签和列名
df.iat[0, 1] = 26 # 使用行和列的整数索引
print(df) Age City AgeSquared Category New_Column Age_Mean_By_City \
Name
ALICE 26.0 26 676.0 New York 30 or under 26.0
BOB 31.0 Paris 961.0 Paris Over 30 31.0
CHARLIE 36.0 London 1296.0 London Over 30 33.5
A 31.0 London 961.0 London Over 30 33.5
0 26.0 NaN NaN NaN NaN NaN Age_Rank Name_Length Name_Age Split_Name Split_Age
Name
ALICE 1.0 8.0 ALICE_26.0 ALICE 26.0
BOB 2.0 5.0 BOB_31.0 BOB 31.0
CHARLIE 3.0 6.0 CHARLIE_36.0 CHARLIE 36.0
A 2.0 6.0 A_31.0 A 31.0
0 NaN NaN NaN NaN NaN
# 使用 cut 和 qcut 进行数据分箱
# 数据分箱
df['Age_Bin'] = pd.cut(df['Age'], bins=[20, 30, 40, 50], labels=["20s", "30s", "40s"])
# 使用 pandas.cut 进行分箱，同时去除重复边界
df['Age_Quantile'] = pd.qcut(df['Age'], q=4, labels=False, duplicates='drop')
print(df) Age City AgeSquared Category New_Column Age_Mean_By_City \
Name
ALICE 26.0 26 676.0 New York 30 or under 26.0
BOB 31.0 Paris 961.0 Paris Over 30 31.0
CHARLIE 36.0 London 1296.0 London Over 30 33.5
A 31.0 London 961.0 London Over 30 33.5
0 26.0 NaN NaN NaN NaN NaN Age_Rank Name_Length Name_Age Split_Name Split_Age Age_Bin \
Name
ALICE 1.0 8.0 ALICE_26.0 ALICE 26.0 20s
BOB 2.0 5.0 BOB_31.0 BOB 31.0 30s
CHARLIE 3.0 6.0 CHARLIE_36.0 CHARLIE 36.0 30s
A 2.0 6.0 A_31.0 A 31.0 30s
0 NaN NaN NaN NaN NaN 20s Date Year Month Day City_Codes Age_Quantile
Name
ALICE 2020-01-27 2020 1 27 0 0
BOB 2020-02-01 2020 2 1 2 0
CHARLIE 2020-02-06 2020 2 6 1 1
A 2020-02-01 2020 2 1 1 0
0 2020-01-27 2020 1 27 -1 0
# 处理日期和时间数据
# 处理日期和时间数据
df['Date'] = pd.to_datetime('2020-01-01') + pd.to_timedelta(df['Age'], unit='D')
df['Year'] = df['Date'].dt.year
df['Month'] = df['Date'].dt.month
df['Day'] = df['Date'].dt.day
print(df) Age City AgeSquared Category New_Column Age_Mean_By_City \
Name
ALICE 26.0 26 676.0 New York 30 or under 26.0
BOB 31.0 Paris 961.0 Paris Over 30 31.0
CHARLIE 36.0 London 1296.0 London Over 30 33.5
A 31.0 London 961.0 London Over 30 33.5
0 26.0 NaN NaN NaN NaN NaN Age_Rank Name_Length Name_Age Split_Name Split_Age Age_Bin \
Name
ALICE 1.0 8.0 ALICE_26.0 ALICE 26.0 20s
BOB 2.0 5.0 BOB_31.0 BOB 31.0 30s
CHARLIE 3.0 6.0 CHARLIE_36.0 CHARLIE 36.0 30s
A 2.0 6.0 A_31.0 A 31.0 30s
0 NaN NaN NaN NaN NaN 20s Date Year Month Day
Name
ALICE 2020-01-27 2020 1 27
BOB 2020-02-01 2020 2 1
CHARLIE 2020-02-06 2020 2 6
A 2020-02-01 2020 2 1
0 2020-01-27 2020 1 27
# 分类数据类型的操作
# 处理分类数据类型
df['City'] = df['City'].astype('category')
df['City_Codes'] = df['City'].cat.codes
print(df) Age City AgeSquared Category New_Column Age_Mean_By_City \
Name
ALICE 26.0 26 676.0 New York 30 or under 26.0
BOB 31.0 Paris 961.0 Paris Over 30 31.0
CHARLIE 36.0 London 1296.0 London Over 30 33.5
A 31.0 London 961.0 London Over 30 33.5
0 26.0 NaN NaN NaN NaN NaN Age_Rank Name_Length Name_Age Split_Name Split_Age Age_Bin \
Name
ALICE 1.0 8.0 ALICE_26.0 ALICE 26.0 20s
BOB 2.0 5.0 BOB_31.0 BOB 31.0 30s
CHARLIE 3.0 6.0 CHARLIE_36.0 CHARLIE 36.0 30s
A 2.0 6.0 A_31.0 A 31.0 30s
0 NaN NaN NaN NaN NaN 20s Date Year Month Day City_Codes
Name
ALICE 2020-01-27 2020 1 27 0
BOB 2020-02-01 2020 2 1 2
CHARLIE 2020-02-06 2020 2 6 1
A 2020-02-01 2020 2 1 1
0 2020-01-27 2020 1 27 -1
# 缺失数据的插值
df_with_na = df.copy()
# 重置索引，将当前索引转移到列
df_with_na = df_with_na.reset_index()
# 如果您希望 'Name' 列的名称是 'Name' 而不是原索引的名称
df_with_na = df_with_na.rename(columns={'Name': 'Name'})
df_with_na.loc[1:3, 'Age'] = np.nan
print(df_with_na)
df_with_na['Age'] = df_with_na['Age'].interpolate()
print(df_with_na) Name Age City AgeSquared Category New_Column Age_Mean_By_City \
0 ALICE 26.0 26 676.0 New York 30 or under 26.0
1 BOB NaN Paris 961.0 Paris Over 30 31.0
2 CHARLIE NaN London 1296.0 London Over 30 33.5
3 A NaN London 961.0 London Over 30 33.5
4 0 26.0 NaN NaN NaN NaN NaN Age_Rank Name_Length Name_Age Split_Name Split_Age Age_Bin \
0 1.0 8.0 ALICE_26.0 ALICE 26.0 20s
1 2.0 5.0 BOB_31.0 BOB 31.0 30s
2 3.0 6.0 CHARLIE_36.0 CHARLIE 36.0 30s
3 2.0 6.0 A_31.0 A 31.0 30s
4 NaN NaN NaN NaN NaN 20s Date Year Month Day City_Codes Age_Quantile
0 2020-01-27 2020 1 27 0 0
1 2020-02-01 2020 2 1 2 0
2 2020-02-06 2020 2 6 1 1
3 2020-02-01 2020 2 1 1 0
4 2020-01-27 2020 1 27 -1 0 Name Age City AgeSquared Category New_Column Age_Mean_By_City \
0 ALICE 26.0 26 676.0 New York 30 or under 26.0
1 BOB 26.0 Paris 961.0 Paris Over 30 31.0
2 CHARLIE 26.0 London 1296.0 London Over 30 33.5
3 A 26.0 London 961.0 London Over 30 33.5
4 0 26.0 NaN NaN NaN NaN NaN Age_Rank Name_Length Name_Age Split_Name Split_Age Age_Bin \
0 1.0 8.0 ALICE_26.0 ALICE 26.0 20s
1 2.0 5.0 BOB_31.0 BOB 31.0 30s
2 3.0 6.0 CHARLIE_36.0 CHARLIE 36.0 30s
3 2.0 6.0 A_31.0 A 31.0 30s
4 NaN NaN NaN NaN NaN 20s Date Year Month Day City_Codes Age_Quantile
0 2020-01-27 2020 1 27 0 0
1 2020-02-01 2020 2 1 2 0
2 2020-02-06 2020 2 6 1 1
3 2020-02-01 2020 2 1 1 0
4 2020-01-27 2020 1 27 -1 0
# 数据帧的压缩和展开（堆叠与反堆叠）
# 堆叠与反堆叠
stacked_df = df.stack()
unstacked_df = stacked_df.unstack()
print(stacked_df)
print(unstacked_df)Name
ALICE Age 26.0City 26AgeSquared 676.0Category New YorkNew_Column 30 or under...
0 Date 2020-01-27 00:00:00Year 2020Month 1Day 27City_Codes -1
Length: 75, dtype: objectAge City AgeSquared Category New_Column Age_Mean_By_City \
Name
ALICE 26.0 26 676.0 New York 30 or under 26.0
BOB 31.0 Paris 961.0 Paris Over 30 31.0
CHARLIE 36.0 London 1296.0 London Over 30 33.5
A 31.0 London 961.0 London Over 30 33.5
0 26.0 NaN NaN NaN NaN NaN Age_Rank Name_Length Name_Age Split_Name Split_Age Age_Bin \
Name
ALICE 1.0 8.0 ALICE_26.0 ALICE 26.0 20s
BOB 2.0 5.0 BOB_31.0 BOB 31.0 30s
CHARLIE 3.0 6.0 CHARLIE_36.0 CHARLIE 36.0 30s
A 2.0 6.0 A_31.0 A 31.0 30s
0 NaN NaN NaN NaN NaN 20s Date Year Month Day City_Codes
Name
ALICE 2020-01-27 00:00:00 2020 1 27 0
BOB 2020-02-01 00:00:00 2020 2 1 2
CHARLIE 2020-02-06 00:00:00 2020 2 6 1
A 2020-02-01 00:00:00 2020 2 1 1
0 2020-01-27 00:00:00 2020 1 27 -1
NumPy 示例
NumPy 是一个用于数值计算的库，特别适合处理大型多维 数组和矩阵。
# 创建和操作数组
import numpy as np# 创建一个一维数组
arr = np.array([1, 2, 3, 4, 5])# 打印数组和它的形状
print("Array:", arr)
print("Shape:", arr.shape)Array: [1 2 3 4 5]
Shape: (5,)
# 创建一个二维数组
arr_2d = np.array([[1, 2, 3], [4, 5, 6]])# 打印数组和它的形状
print("2D Array:\n", arr_2d)
print("Shape:", arr_2d.shape)2D Array:[[1 2 3][4 5 6]]
Shape: (2, 3)
# 创建一个数组
arr = np.array([1, 2, 3, 4, 5])# 数组乘法
arr_times_two = arr * 2# 显示结果
print("Array multiplied by 2:", arr_times_two)Array multiplied by 2: [ 2 4 6 8 10]
# 创建零数组
# 创建一个形状为 3x3 的零数组
zeros = np.zeros((3, 3))
print(zeros)[[0. 0. 0.][0. 0. 0.][0. 0. 0.]]
# 创建单位矩阵
# 创建一个 3x3 单位矩阵
identity_matrix = np.eye(3)
print(identity_matrix)[[1. 0. 0.][0. 1. 0.][0. 0. 1.]]
# 创建等间隔数值数组
# 创建一个从 0 到 10 的等间隔的数组
linspace_arr = np.linspace(0, 10, 5)
print(linspace_arr)[ 0. 2.5 5. 7.5 10. ]
# 数组形状更改
# 更改数组的形状
# 创建一个简单的一维数组
arr1d = np.array([1, 2, 3, 4, 5])
reshaped_arr = np.reshape(arr1d, (5, 1))
print(reshaped_arr)[[1][2][3][4][5]]
# 数组索引
# 创建一个二维数组
arr2d = np.array([[1, 2, 3], [4, 5, 6]])
# 索引一个二维数组
element = arr2d[0, 1] # 获取第一行第二列的元素
print(element)2
# 数组切片
# 切片数组
slice_arr = arr2d[0:2, 1:3]
print(slice_arr)[[2 3][5 6]]
# 数组数据类型转换
# 转换数组的数据类型
float_arr = arr1d.astype(float)
print(float_arr)[1. 2. 3. 4. 5.]
# 数组堆叠
# 水平堆叠两个数组
hstack_arr = np.hstack((arr1d, arr1d))
print(hstack_arr)[1 2 3 4 5 1 2 3 4 5]
# 数组分割
# 水平分割数组
split_arr = np.hsplit(arr2d, 3)
print(split_arr)[array([[1],[4]]), array([[2],[5]]), array([[3],[6]])]
# 数组迭代
# 迭代二维数组
for row in arr2d:print(row)[1 2 3]
[4 5 6]
# 数组条件筛选
# 使用条件筛选数组
filtered_arr = arr1d[arr1d > 3]
print(filtered_arr)[4 5]
# 数组元素的数学操作
# 对数组元素执行数学操作
squared_arr = np.square(arr1d)
print(squared_arr)[ 1 4 9 16 25]
# 数组统计计算
# 数组的统计计算
mean_value = np.mean(arr1d)
print(mean_value)3.0
# 数组的广播机制
# 使用广播机制执行操作
broadcast_arr = arr1d + 2
print(broadcast_arr)[3 4 5 6 7]
# 数组的线性代数运算
# 矩阵乘法
matrix_product = np.dot(arr2d, arr2d.T)
print(matrix_product)[[14 32][32 77]]
# 数组的排序
# 对数组进行排序
sorted_arr = np.sort(arr1d)
print(sorted_arr)[1 2 3 4 5]
# 查找唯一元素和计数
# 查找数组中的唯一元素和它们的计数
unique_elements, counts = np.unique(arr1d, return_counts=True)
print(unique_elements, counts)[1 2 3 4 5] [1 1 1 1 1]
# 数组的文件操作
# 将数组保存到文件
np.save('array.npy', arr1d)# 从文件加载数组
loaded_arr = np.load('array.npy')
print(loaded_arr)[1 2 3 4 5]
# 随机数生成
# 生成随机数数组
random_arr = np.random.rand(5)
print(random_arr)[0.28233229 0.48873143 0.63405945 0.67577065 0.31475481]
# 数组的累积操作
# 数组的累积求和
cumsum_arr = np.cumsum(arr1d)
print(cumsum_arr)[ 1 3 6 10 15]
# 数组的逻辑操作
# 数组的逻辑运算
logical_arr = np.logical_and(arr1d > 2, arr1d < 5)
print(logical_arr)[False False True True False]
# 数组的复数操作
# 创建复数数组
complex_arr = np.array([1+2j, 3+4j])
print(complex_arr.real) # 实部
print(complex_arr.imag) # 虚部[1. 3.]
[2. 4.]
# 数组的梯度计算
# 计算数组的梯度
gradient_arr = np.gradient(arr1d)
print(gradient_arr)[1. 1. 1. 1. 1.]
# 数组的标准差和方差计算
# 计算标准差和方差
std_dev = np.std(arr1d)
variance = np.var(arr1d)
print(std_dev, variance)1.4142135623730951 2.0
# 数组的卷积运算
# 计算两个数组的卷积
conv_arr = np.convolve([1, 2, 3], [0, 1, 0.5])
print(conv_arr)[0. 1. 2.5 4. 1.5]
# 创建结构化数组
# 创建结构化数组
structured_arr = np.array([('Alice', 25), ('Bob', 30)], dtype=[('Name', 'U10'), ('Age', 'i4')])
print(structured_arr)[('Alice', 25) ('Bob', 30)]
# 数组的内存布局
# 查看数组的内存布局
print(arr1d.flags) C_CONTIGUOUS : TrueF_CONTIGUOUS : TrueOWNDATA : TrueWRITEABLE : TrueALIGNED : TrueWRITEBACKIFCOPY : False
# 高级索引技巧
# 使用高级索引技巧
fancy_indexed_arr = arr2d[[0, 1], [1, 2]]
print(fancy_indexed_arr)[2 6]
数据分析案例: Iris 数据集
# 获取数据
import pandas as pd
import os# Iris 数据集的 URL
url = "https://raw.githubusercontent.com/uiuc-cse/data-fa14/gh-pages/data/iris.csv"
if os.path.exists('iris.csv'):# 读取数据集iris = pd.read_csv(url)
else:# 读取数据集iris = pd.read_csv(url)iris.to_csv('iris.csv', index=False)
# 显示前几行
print(iris.head()) sepal_length sepal_width petal_length petal_width species
0 5.1 3.5 1.4 0.2 setosa
1 4.9 3.0 1.4 0.2 setosa
2 4.7 3.2 1.3 0.2 setosa
3 4.6 3.1 1.5 0.2 setosa
4 5.0 3.6 1.4 0.2 setosa
# 数据探索
# 数据集的基本信息
print(iris.info())# 统计描述
print(iris.describe())# 检查缺失值
print(iris.isnull().sum())<class 'pandas.core.frame.DataFrame'>
RangeIndex: 150 entries, 0 to 149
Data columns (total 5 columns):# Column Non-Null Count Dtype
--- ------ -------------- ----- 0 sepal_length 150 non-null float641 sepal_width 150 non-null float642 petal_length 150 non-null float643 petal_width 150 non-null float644 species 150 non-null object
dtypes: float64(4), object(1)
memory usage: 6.0+ KB
Nonesepal_length sepal_width petal_length petal_width
count 150.000000 150.000000 150.000000 150.000000
mean 5.843333 3.054000 3.758667 1.198667
std 0.828066 0.433594 1.764420 0.763161
min 4.300000 2.000000 1.000000 0.100000
25% 5.100000 2.800000 1.600000 0.300000
50% 5.800000 3.000000 4.350000 1.300000
75% 6.400000 3.300000 5.100000 1.800000
max 7.900000 4.400000 6.900000 2.500000
sepal_length 0
sepal_width 0
petal_length 0
petal_width 0
species 0
dtype: int64
# 比较两个特征，萼片长度（sepal length）和萼片宽度（sepal width），并根据鸢尾花的种类对点进行着色。
import matplotlib.pyplot as plt# 为不同的物种设置不同的颜色
species_colors = {'setosa': 'red', 'versicolor': 'green', 'virginica': 'blue'}# 创建散点图
plt.figure(figsize=(8, 6))
for species, color in species_colors.items():# 选择特定物种的数据species_data = iris[iris['species'] == species]plt.scatter(species_data['sepal_length'], species_data['sepal_width'],color=color, label=species)# 添加标题和标签
plt.title('Iris Sepal Measurements')
plt.xlabel('Sepal Length (cm)')
plt.ylabel('Sepal Width (cm)')# 显示图例
plt.legend()# 显示图表
plt.show()
#数据处理
# 创建一个新列，为花瓣面积
iris['petal_area'] = iris['petal_length'] * iris['petal_width']# 显示新的数据集头部
print(iris.head()) sepal_length sepal_width petal_length petal_width species petal_area
0 5.1 3.5 1.4 0.2 setosa 0.28
1 4.9 3.0 1.4 0.2 setosa 0.28
2 4.7 3.2 1.3 0.2 setosa 0.26
3 4.6 3.1 1.5 0.2 setosa 0.30
4 5.0 3.6 1.4 0.2 setosa 0.28
# 数据分析
# 按种类分组并计算平均值
print(iris.groupby('species').mean()) sepal_length sepal_width petal_length petal_width petal_area
species
setosa 5.006 3.418 1.464 0.244 0.3628
versicolor 5.936 2.770 4.260 1.326 5.7204
virginica 6.588 2.974 5.552 2.026 11.2962
