Time-Series Metabolomic Data Aggregation Pipeline

Create a robust PostgreSQL stored procedure that processes time-series metabolomic experimental data, handling multiple instrument readings with different sampling frequencies. The procedure must: 1) Normalize raw spectral data across multiple experimental batches, 2) Handle missing data points with intelligent interpolation, 3) Calculate rolling statistical windows for compound concentration changes, 4) Generate a final aggregated result set with confidence intervals and metadata tracking.