Scenario

A sales dataset contains missing values across different column types that need to be cleaned before analysis.

Task

Write a Python script at /home/interview/clean_sales.py using pandas that reads /home/interview/sales_data.csv, replaces null values based on column type (numeric columns with 0, text columns with "Unknown"), and saves the cleaned dataset to /home/interview/cleaned_data.csv.

Example

Input with nulls:

order_id,customer_name,quantity,status
1,John Doe,5,completed
2,,NaN,
3,Jane Smith,10,pending

Expected output:

order_id,customer_name,quantity,status
1,John Doe,5,completed
2,Unknown,0,Unknown
3,Jane Smith,10,pending

Step 1: Examine the dataset

head /home/interview/sales_data.csv

Review the columns and identify which are numeric and which are text-based.

Step 2: Create the Python script

nano /home/interview/clean_sales.py

Write a script to replace nulls based on column type:

import pandas as pd

# Read the CSV file
df = pd.read_csv('/home/interview/sales_data.csv')

# Identify numeric and text columns
numeric_cols = df.select_dtypes(include=['number']).columns
text_cols = df.select_dtypes(include=['object', 'string']).columns

# Replace nulls in numeric columns with 0
df[numeric_cols] = df[numeric_cols].fillna(0)

# Replace nulls in text columns with "Unknown"
df[text_cols] = df[text_cols].fillna("Unknown")

# Save cleaned dataset
df.to_csv('/home/interview/cleaned_data.csv', index=False)

print(f"Cleaned {len(df)} records")
print(f"Numeric columns filled with 0: {list(numeric_cols)}")
print(f"Text columns filled with 'Unknown': {list(text_cols)}")

The select_dtypes() method identifies columns by data type, then fillna() replaces nulls with appropriate values.

Step 3: Run the script

python3 /home/interview/clean_sales.py

Step 4: Verify the output

head /home/interview/cleaned_data.csv

Should show no null values, with 0 in numeric columns and "Unknown" in text columns where data was missing.