TypeError:无法根据规则“safe”将数组数据从dtype（'<U1'）强制转换为dtype（'float64'）

Question

当我试图在散点图上使用不同颜色绘制异常值时，遇到了以下错误：

TypeError: Cannot cast array data from dtype('U1') to dtype('float64') according to the rule 'safe'

我的代码：

import statsmodels.api as sm
import pandas as pd
import matplotlib.pyplot as plt
import numpy as np

CRP = pd.read_csv('CarsProp.csv')

priceMean = CRP.price.mean()
priceStd = CRP.price.std()   

CRP['isOutlierPrice'] = np.nan
testColumn1 = abs(CRP.price - priceMean) > 2*priceStd

for i, value in enumerate(testColumn1):
    if value == True:
        CRP['isOutlierPrice'][i] = 1

mileageMean = CRP.mileage.mean()
mileageStd = CRP.mileage.mean()

CRP['isOutlierMileage'] = np.nan
testColumn2 = abs(CRP.mileage - mileageMean) > 2*priceStd

for i, value in enumerate(testColumn2):
    if value == True:
        CRP['isOutlierMileage'][i] = 1

outlierPmsJoint = ((CRP['isOutlierPrice'] == 1) | (CRP['isOutlierMileage'] == 1))

colorChoiceDict = {True: (1.0, 0.55, 0.0, 1.0), 
              False: (0.11, 0.65, 0.72, 0.1)}

colorCol = [colorChoiceDict[val] for val in outlierPmsJoint]

PriceFloat = [float(val) for val in CRP.price]
MileageFloat = [float(val) for val in CRP.mileage]

plt.figure()
plt.scatter(PriceFloat, MileageFloat, c = colorCol, linewidth='0')
plt.set_title('Price vs. Mileage with outliers')  

有人知道问题在哪里，怎么解决吗？谢谢你

Answer 1

问题出在散点图上。它应该是：

plt.scatter(PriceFloat, MileageFloat, c = colorCol, linewidth=0)

我把行宽度arg放在倒转的comas中，但没有产生输出。然而，错误信息非常误导人。下一次调试时我可能不会太注意错误消息。在