Pandas Missing Value Imputation Pipelineeasy

Pandas Missing Value Imputation Pipeline

easy NumPy & Data Wrangling 10–15 min

NumPy allowed · plain Python earns 10 bonus points

A vector transforms

Matrix multiplication is the first computation behind every model.

Open lesson →

01vector

02matrix

03transform

04result

[2, −1] × [[1, 2], [−.5, 1]] = [0, −2]

Computed matrix × vector transformation

Impute missing values across numerical and categorical features in tabular records using Pandas.

Related Skills

Data Pipelines & ETL

Task

Implement impute_missing_values(records, num_strategy, cat_strategy). For numeric columns with missing values, fill with the column mean or median (rounded to 2 decimal places). For non-numeric columns, fill with the mode or string fallback.

Function Signature

impute_missing_values(records: list[dict], num_strategy: str = 'mean', cat_strategy: str = 'mode') -> list[dict]

Examples

Example 1: Numeric Mean and Categorical Mode
Input: records = [{'age': 20, 'city': 'NY'}, {'age': 40, 'city': 'SF'}, {'age': None, 'city': 'NY'}, {'age': 30, 'city': None}]
num_strategy = 'mean'
cat_strategy = 'mode'
Output: [{'age': 20, 'city': 'NY'}, {'age': 40, 'city': 'SF'}, {'age': 30, 'city': 'NY'}, {'age': 30, 'city': 'NY'}]

Example 2: Numeric Median and Categorical Constant
Input: records = [{'score': 10, 'grade': 'A'}, {'score': 20, 'grade': None}, {'score': 90, 'grade': 'B'}]
num_strategy = 'median'
cat_strategy = 'missing'
Output: [{'score': 10, 'grade': 'A'}, {'score': 20, 'grade': 'missing'}, {'score': 90, 'grade': 'B'}]

Constraints

records is a list of dictionary rows.
num_strategy is either 'mean' or 'median'.
cat_strategy is 'mode' or a string replacement.
Returns a list of dicts in identical row order with all missing values imputed.

Python3Saved ✓

import pandas as pd

import numpy as np

def impute_missing_values(records: list[dict], num_strategy: str = 'mean', cat_strategy: str = 'mode') -> list[dict]:

"""Impute missing (None/NaN) values in records using pandas.

For numeric columns: fill with mean or median (rounded to 2 decimal places).

For categorical columns: fill with mode or literal placeholder.

Return the cleaned records as list of dicts."""

# Your implementation here

pass

Case 1: Numeric Mean and Categorical Mode

Input: records = [{'age': 20, 'city': 'NY'}, {'age': 40, 'city': 'SF'}, {'age': None, 'city': 'NY'}, {'age': 30, 'city': None}]
num_strategy = 'mean'
cat_strategy = 'mode'

Expected: [{'age': 20, 'city': 'NY'}, {'age': 40, 'city': 'SF'}, {'age': 30, 'city': 'NY'}, {'age': 30, 'city': 'NY'}]

Case 2: Numeric Median and Categorical Constant

Input: records = [{'score': 10, 'grade': 'A'}, {'score': 20, 'grade': None}, {'score': 90, 'grade': 'B'}]
num_strategy = 'median'
cat_strategy = 'missing'

Expected: [{'score': 10, 'grade': 'A'}, {'score': 20, 'grade': 'missing'}, {'score': 90, 'grade': 'B'}]