• Data Cleaning, Parsing & Standardization (Core Function) • Build Python (Pandas/Regex) workflows to extract structure from free-form Prisma text fields containing placements, publishers, formats, devices, tactics, etc. • Develop rule-based and pattern-based logic to recognize abbreviation variations, naming inconsistencies, and agency-specific formatting habits. • Create normalization scripts that standardize publisher names, platform types, channels, and other key taxonomies. • Maintain “classification dictionaries” and pattern libraries that improve accuracy over time. • Data Quality & Operational Support • Profile incoming agency data to identify common quality issues and propose concrete remediation. • Design QA checks to detect missing fields, malformed entries, inconsistent dimensional labels, and mismatched spend totals. • Partner closely with Data Engineering to embed your logic into scheduled pipelines. • Own recurring operational workflows that support data ingestion, aggregation, and quality monitoring. • Dimensional Enrichment, Modeling & Data Derivation • Build and maintain predictive models to support media spend forecasting, pricing analysis, and text classification for data tagging and enrichment. • Apply statistical modeling or regression techniques to impute missing values or infer incomplete spend or performance metrics. • Derive new fields—such as format groupings, platform classifications, creative types, and targeting attributes—using pattern recognition and model-based inference across text fields. • Produce structured dimensions from unstructured data to improve downstream reporting. • Document the logic and assumptions behind model-driven derivations to ensure transparency and reproducibility. • Analysis & Internal Stakeholder Support • Investigate anomalies, unexpected trends, and data gaps by tracing issues back to source fields. • Build dashboards that help internal teams observe data quality, volume, deviations, and model-driven insights. • Provide ad-hoc deep dives into digital spend patterns, forecast accuracy, and classification performance for product, ops, and leadership.