Pandas模块 -- 数据类型转换，描述统计

car=pd.read_csv(r'E:\Python\sec_cars.csv',sep=',').head(32)
# print(car)
print("数据集的类型：",type(car))
print("数据集的维数：",car.ndim)
print("数据集的行列数：",car.shape)
print("数据集car的元素类型:\n",car.dtypes)

car.Boarding_time=pd.to_datetime(car.Boarding_time,format='%Y年%m月')
car.New_price=car.New_price.str[:-1].astype("float")
print("修改某些字段的数据类型后：\n",car.dtypes)
#计算数值型变量的统计描述
print("数值型变量的统计描述:\n",car.describe())

#计算数值型变量的偏度和峰度
num=car.columns[car.dtypes!='object'][1:]

def skew_kurt(x):
    skewness=x.skew()
    kurtsis=x.kurt()
    return pd.Series([skewness,kurtsis],index=['skewness','kurtsis'])

skew_kurt=car[num].apply(func=skew_kurt,axis=0)
print("数值型变量的偏度和峰度:\n",skew_kurt)

#计算离散型变量的统计描述
print("离散型变量的统计描述:\n",car.describe(include=['object']))

count=car.Discharge.value_counts()
ratio=count/car.shape[0]
c=pd.DataFrame({'freq':count,'freq_ratio':ratio})
print("离散型变量的频次和频率:\n",c)
print("")
print("重置行索引后的离散型变量的频次和频率：")
c.reset_index(inplace=True)
print(c)