I'm learning object oriented programing in a data science context.

I want to understand what good practice is in terms of writing methods within a class that relate to one another.

When I run my code:

import pandas as pd 
pd.options.mode.chained_assignment = None  

class MyData:
    def __init__(self, file_path):
        self.file_path = file_path

def prepper_fun(self):
    '''Reads in an excel sheet, gets rid of missing values and sets datatype to numerical'''
    df = pd.read_excel(self.file_path) 
    df = df.dropna()                    
    df = df.apply(pd.to_numeric) 
    self.df = df
    return(df)

def quality_fun(self):
   '''Checks if any value in any column is more than 10. If it is, the value is replaced with
   a warning 'check the original data value'.'''
    for col in self.df.columns:
        for row in self.df.index:                                             
            if self.df[col][row] > 10:       
                self.df[col][row] = str('check original data value') 
    return(self.df) 

data = MyData('https://archive.ics.uci.edu/ml/machine-learning-databases/00429/Cryotherapy.xlsx')
print(data.prepper_fun())
print(data.quality_fun())

I get the following output (only part of the output is shown due to space constrains):

    sex  age   Time  
0     1   35  12.00             
1     1   29   7.00               
2     1   50   8.00                
3     1   32  11.75                
4     1   67   9.25                         
..  ...  ...    ...       


     sex                        age                       Time 
0     1  check original data value  check original data value                  
1     1  check original data value                          7                  
2     1  check original data value                          8                  
3     1  check original data value  check original data value               
4     1  check original data value                       9.25 
..  ...                        ...                        ...

I am happy with the output generated by each method.

But if I try to call print(data.quality_fun()) without first calling print(data.prepper_fun()), I get an error AttributeError: 'MyData' object has no attribute 'df'.

Being new to objected oriented programming, I am wondering if it is considered good practice to structure things like this, or if there is some other way of doing it.

Thanks for any help!

class feed(object):

    # other functions

    def new_method(self):
        #This can now reference other object attributes​

class feed(object):

    def __init__ (self, food = 0, water = 0) :
        self.food = food
        self.water = water

    def fill(self): 
        self.food += 1
        self.water += 1​

Make sure you have the df before you use it.

class MyData:
    def __init__(self, file_path):
        self.file_path = file_path
        self.df = None
    def quality_fun():
       if self.df is None:
          self.prepper_fun()
       # rest of the code 

If the csv file isn't changed during runtime you should call prepper_fun(self)  in the __init__. , calling it separately leads to a high chance of bugs.

If the csv file is changed then the other answer works perfectly well