[lab_customer_analysis_round1] leonardo_pagliacci

[leonardo-pagliacci]

Lab 1 - Customer Analysis Round 1 - Leonardo Pagliacci solutions.

[pt-data-ta-lis]

Good Job!

Comments:

• Rearrange the columns in the dataframe as needed
  cus_data2 = cus_data2[cus_data1.columns] # Changing the order of the columns
  cus_data3 = cus_data3.rename(columns={'state':'st'}) # renaming the column 'state'
  cus_data3 = cus_data3[cus_data1.columns] # Changing the order of the columns

• Which columns are numerical?
  data.select_dtypes('float')

• Which columns are categorial?
  data.select_dtypes('object')

• Delete the column education, a number_of_open_complaints from the dataframe
  data = data.drop(['education', 'number_of_open_complaints'], axis=1)

• Correct the values in the column customer_lifetime_value. They are given as percent, so multiply them by 100 and change dtype to numerical type
  cus_data['customer_lifetime_value'] = cus_data['customer_lifetime_value'].astype(str) --> This one line isn't useful, if you look at the dtype of this column is "object" so it's either a str or NaN and NaN values (Not a Number) cannot be directly converted to strings because they are a specific type of value used to represent an error condition or an undefined result. Trying to convert NaN directly into a string would not provide meaningful information about the error or the context in which it occurred.
  data['customer_lifetime_value'] = data['customer_lifetime_value']*100
  data['customer_lifetime_value'] = pd.to_numeric(data['customer_lifetime_value'], errors='coerce')