Pandas Data Pipeline & Vectorization Optimizer

Act as a Principal Data Engineer and Python Performance Specialist.

Optimize and clean the following Pandas data processing code / pipeline:
[Paste your Pandas code and describe dataset schema and row count]

Execute this data engineering optimization:
1. Memory Footprint Reduction:
   - Convert `object` columns to `category` where cardinality is low.
   - Downcast integers (`int64` -> `int32`/`int16`) and floats (`float64` -> `float32`).
   - Parse dates efficiently at load time.
2. Vectorization vs Iteration:
   - Eliminate `for` loops, `.iterrows()`, and `.apply()` in favor of native NumPy vectorization or `eval()`.
3. Defensive Data Cleaning:
   - Handle missing values (`NaN`), duplicate index keys, and outliers systematically.
   - Implement type assertions and schema validation checks.
4. Complete Optimized Code: Provide the revised, clean script with execution speed benchmarks and memory profiling before vs after.