Objective:

The objective of this project is to read data from a CSV file, perform grouping and sampling operations based on specified attributes, and create a new DataFrame with the sampled data.

Pseudocode:

1. Read data from Data.csv into DataFrame df0.
2. Group df0 by the attributes ["Sex", "Age_category", "Highest_education_level"] to obtain the frequency of each combination, storing the result in a dictionary freq_of_all_combos.
3. Initialise an empty DataFrame df with 50,000 rows and columns ["Sex", "Age_category", "Highest_education_level"].
4. For each unique combination combo in freq_of_all_combos:
   • Calculate the proportion of combo in the seed sample as proportion = (frequency of the combination / total sample size).
   • Determine the number of agents n in df corresponding to combo using the formula n = proportion * 50000.
   • Identify the indices in df where the "Sex" column is NaN and store them in sample_indices.
   • If the number of sample_indices is greater than or equal to n:
     • Randomly select n indices from sample_indices without replacement.
     • Assign the values of combo to the selected indices in df.
5. Return the df DataFrame.