This lesson's interactive features are locked, please to keep using them

Validation Summary

If there isn't a trustworthy fix for invalid rows, sometimes it's best just to remove them:

valid_mask = df["rating"].between(1, 5)
df_clean = df[valid_mask]

That said, silently dropping bad rows can hide real problems. I like to add a validation summary so that the damage remains visible:

invalid_mask = ~df["rating"].between(1, 5)
validation_summary = {
    "total_rows": len(df),
    "invalid_ratings": invalid_mask.sum(),
}

Assignment

Complete the remove_invalid_device_data function. It accepts a DataFrame with temperature and battery_level columns. Return a cleaned DataFrame with only valid records and a validation summary dictionary.

Create a copy of the DataFrame to avoid modifying the original.
Build a valid mask where temperature is between -40 and 200, inclusive.
Build a valid mask where battery_level is between 0 and 100, inclusive.
Build a validation summary dictionary with these keys: total_rows and invalid_rows.
Create a new DataFrame with only the valid rows.
Reset the index, dropping the old index numbers.
Return a tuple: the cleaned DataFrame, then the validation summary dictionary.