Python Script for Normalizing Multi-Source Property Listing Feeds

Role: Senior Real Estate Data Engineer specializing in MLS aggregation and listing normalization.

Context

• Raw Listing Feed Source: {{raw_listing_file_path}}
• Target Schema Mapping: {{target_schema_mapping}}
• Deduplication Similarity Threshold: {{duplicate_threshold}}
• Metric Imputation Strategy: {{imputation_strategy}}
• Clean Output File Path: {{output_cleaned_path}}

Task

Write a fully executable Python data cleaning script using pandas and regex to ingest unstandardized real estate listing data from {{raw_listing_file_path}}, rectify malformed addresses, standardize inconsistent property categories against {{target_schema_mapping}}, resolve duplicates using {{duplicate_threshold}}, and save the sanitized dataset to {{output_cleaned_path}}.

Method

1. Ingest the raw data from {{raw_listing_file_path}} into a pandas DataFrame, inferring and logging data types.
2. Parse and normalize mixed address strings into standard fields (street number, street name, unit, city, state, postal code) using regular expressions.
3. Map heterogeneous property type strings (e.g., 'TH', 'Townhouse', 'Row Home') to standardized schema values defined in {{target_schema_mapping}}.
4. Sanitize numerical attributes including price, lot acreage, and interior square footage by removing non-numeric characters and converting to numeric floats.
5. Handle missing numerical attributes using {{imputation_strategy}} grouped by geographic zip code and property category.
6. Identify and merge cross-listed duplicate records by applying string similarity matching on normalized address and price using {{duplicate_threshold}}.
7. Format and export the final validated DataFrame to {{output_cleaned_path}} with comprehensive processing statistics printed to stdout.

Constraints

• The script MUST be written in clean, PEP 8 compliant Python 3.10+ using only standard library, pandas, and re.
• The script MUST NOT drop records with missing metrics without logging the exact reason and record ID to a quarantine log.
• Type hints and docstrings MUST be provided for every data cleaning function.
• Avoid hardcoded file paths; use the defined variables.
• Include inline defensive assertions for critical non-null fields.

Output format

Provide a single runnable Python script structured into the following sections:

1. Module docstring and library imports.
2. Configurable constants block mapping to {{raw_listing_file_path}}, {{target_schema_mapping}}, {{duplicate_threshold}}, {{imputation_strategy}}, and {{output_cleaned_path}}.
3. Core modular cleaning functions.
4. A main execution block (if __name__ == '__main__':) demonstrating end-to-end execution with console summary metrics.

Self-review

• Confirm that all 5 variables are used in the generated script logic.
• Verify that regex parsing handles edge cases like suite numbers and directional prefixes.
• Ensure execution handles empty dataframes or schema mismatches gracefully without unhandled exceptions.