How to deal with missing values in data cleaning

The data you inherit for analysis will come from multiple sources and would have been pulled adhoc. So this data will not be immediately ready for you to run any kind of model on. One of the most common issues you will have to deal with is missing values in the dataset. There are many reasons why values might be missing - intentional, user did not fill up, online forms broken, accidentally deleted, legacy issues etc. 

Either way you will need to fix this problem. There are 3 ways to do this - either you will ignore the missing values, delete the missing value rows or fill the missing values with an approximation.

Its easiest to just drop the missing observations but you need to very careful before you do that, because the absence of a value might actually be conveying some information about the data pattern. If you decide to drop missing values :

df_no_missing = df.dropna()

will drop any rows with any value missing. Even if some values are available in a row it will still get dropped even if a single value is missing. 

df_cleaned = df.dropna(how='all')

will only drop rows where all cells are NA or missing values. To drop columns, you will have to add the ‘axis=1’ parameter to the above functions.

The extent of the missing values is identified after identifying the variables with missing values. If any patterns are identified the analyst has to concentrate on them as it could lead to interesting and meaningful business insights. If there are no patterns identified, then the missing values can be substituted with mean or median values (imputation) or they can simply be ignored.There are various factors to be considered when answering this question-

Understand the problem statement, understand the data and then give the answer.Assigning a default value which can be mean, minimum or maximum value. Getting into the data is important.

If it is a categorical variable, the default value is assigned. The missing value is assigned a default value.

If you have a distribution of data coming, for normal distribution give the mean value.

Should we even treat missing values is another important point to consider? If 80% of the values for a variable are missing then you can answer that you would be dropping the variable instead of treating the missing values.