Led data standardization for a Fortune 500 pharmaceutical company in Boston following its merger with a leading biotech firm. The merger had left behind deeply conflicting terminology, business definitions, and incomplete documentation across both entities. The goal was to produce a unified, clean data catalogue and dictionary that could support a reliable database migration going forward.
1. Build the data catalogue and data dictionary
Worked with business, technology, and vendor teams to construct a comprehensive data catalogue. The dictionary was built incrementally — starting with the most critical entities and expanding outward — ensuring buy-in at every stage.
# Example: Conflict-detection between two merged schemas
import pandas as pd
def detect_conflicts(schema_a: pd.DataFrame, schema_b: pd.DataFrame, key="field_name"):
merged = schema_a.merge(schema_b, on=key, how="inner", suffixes=("_a", "_b"))
conflicts = merged[merged["definition_a"] != merged["definition_b"]]
return conflicts[[key, "definition_a", "definition_b"]]
2. Present high-level data flow diagrams for confirmation
Before diving into granular mapping, produced a high-level data flow diagram covering all major systems and handoff points. This was reviewed and signed off by senior stakeholders to lock in the architectural direction.
3. Apply incremental agile approach — one business unit at a time
Rather than a big-bang rollout, adopted an incremental strategy: standardize one business unit and one report at a time, validate, then move to the next. Each iteration proposed an improved solution that linked back to the master structure agreed upon in the earlier phases.
# Example: Incremental validation tracker
def track_unit_status(units: list) -> pd.DataFrame:
return pd.DataFrame(units, columns=[
"business_unit", "fields_total", "fields_standardized",
"reports_migrated", "status" # status: pending | in_progress | validated
])
4. Resolve legacy documentation gaps
Identified and filled documentation gaps inherited from the biotech merger. Created clear field-level descriptions, business rules, and owner assignments so the downstream migration team had a single authoritative reference.
| For more details see LinkedIn | GitHub |