I am preparing a dataset for a model, but somehow the code just doesn't run well.

The major error is:

File "/Users/liangjulia/Desktop/UW DS Certificate Learning Material/untitled6.py", line 61
  'income2' = pd.to_numeric(Adult.income, errors='coerce')
SyntaxError: can't assign to literal


# import statement
import numpy as np
import pandas as pd
import csv
import matplotlib.pyplot as plt
from sklearn.preprocessing import LabelEncoder
# loading dataset, it is a combination of categorical and numerical data
hp = pd.read_csv("https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.data",header=0,sep=',')
hp.columns = ['age','workclass','income','education','education-num','marital-status','occupation','relationship','race','sex','capital-gain','capital-loss','hours-per-week','native-country','salary-range'] 
# dataset basics

# account for all value '?'

#Remove onsolete data point in income

# replace all aberrant values
hp.replace('nan', 0)
hp.replace('NULL', 0)

# change data type of certain data point to numerical
number = LabelEncoder()
hp['income'] = number.fit_transform(hp['income'.astype('str')])
hp['capital-gain'] = number.fit_transform(hp['capital-gain'.astype('str')])
hp['capital-loss'] = number.fit_transform(hp['capital-loss'.astype('str')])

# Choose the datapoint 'income' to perform the data cleaning and remove outliers
LimitHi=np.mean('income') + 2*np.std('income')
LimitLo=np.mean('income') + 2*np.std('income')
BadIncome = ('income' > LimitHi) & ('income' < LimitLo)

# Replace outliars
RightIncome = ~BadIncome
x[BadIncome] = np.mean(x[RightIncome])

# normalize the Income Column using numpy
#'income2' = pd.to_numeric(Adult.income, errors='coerce')
minmaxscaled =('income' - min('income'))/(max('income') - min('income'))

# bin age data into several ranges
hp['bin'] = pd.cut(hp['age'], [15,30,45,60,75,90])

# construct new categorical data point with existing data point 
hp['EvalonInvestment'] = 'zzz'
hp.loc[(hp['capital-gain'] >= 50000), 'loc2'] = 'investmentking'
hp.loc[(hp['capital-gain'] > 10000) & (hp['capital-gain'] < 50000), 'loc2'] = 'good-investment'
hp.loc[(hp['capital-gain'] > 0) & (hp['capital-gain'] <= 10000), 'loc2'] = 'ok-investment'


2 Answers 2


It should be hp['income2'] because you can't assign a mutable object to an immutable object such as a string


What line 61 should read is

income2 = pd.to_numeric(hp['income'], errors = 'coerce')

Let's break this down.

You want to assign a new variable income2 to the output of a method in the pd class, to_numeric.

to_number takes two arguments: arg : list, tuple, 1-d array, or Series, a choice on how to treat errors, and an optional downcast operator. You find this out by


Line 61 originally provided the arg Adult.Income, which could be a list, tuple, 1-d array, or Series, but it hasn't been defined, so it can't be a valid arg.



and you get


which is a valid argument.

This line is not the only problem in your code. I suggest that you use iPython so that you can see all of the other errors. Or just use the command line and copy in a few lines at a time and you'll find new error messages to resolve.

Errors in code are like cockroaches -- they seldom travel alone.

Good luck!


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.