MDM & Cloud Data Warehouse Migration — Tune Protect Insurance, APAC

Led the full master data management (MDM) implementation and on-premise-to-cloud data warehouse migration for Tune Protect, a $260M market-cap insurance company in Southeast Asia. The project spanned underwriting, sales, and marketing functions, requiring deep collaboration between business stakeholders and technology teams to deliver a single source of truth.

1. Build the data dictionary and map data points across applications

Worked with business teams across underwriting, sales, and marketing to define every critical data element. Mapped data points from multiple source applications into a unified schema, resolving naming conflicts and aligning definitions across departments.

# Example: Data dictionary builder
import pandas as pd

def build_data_dictionary(source_schemas: dict) -> pd.DataFrame:
    rows = []
    for source, columns in source_schemas.items():
        for col in columns:
            rows.append({
                "source_system": source,
                "column_name": col["name"],
                "data_type": col["type"],
                "nullable": col.get("nullable", True),
                "description": col.get("desc", ""),
                "mapped_to": col.get("target", "")
            })
    return pd.DataFrame(rows)

2. Execute on-premise to cloud migration

Migrated the entire data warehouse from an in-house on-premise setup to a cloud-based platform. Ensured zero data loss through staged migration with full reconciliation checks at each phase.

3. Deploy CRM-ready analytics in Power BI

After migration, collaborated with function heads to design and deliver tailored Power BI dashboards — daily, weekly, and monthly — for each vertical. The new reporting layer enabled CRM campaigns to achieve over 50% efficacy, a significant uplift from the prior baseline.

# Example: Post-migration reconciliation check
def reconcile(source_df, target_df, key_col, check_cols):
    merged = source_df.merge(target_df, on=key_col, how="outer", suffixes=("_src", "_tgt"), indicator=True)
    mismatches = merged[merged["_merge"] != "both"]
    for col in check_cols:
        diff = merged[merged[f"{col}_src"] != merged[f"{col}_tgt"]]
        print(f"{col}: {len(diff)} mismatches")
    print(f"Unmatched rows: {len(mismatches)}")
    return merged

4. Build conceptual data models and primary key framework

Defined primary keys across all entities, then constructed high-level conceptual data models. These models were tested, deployed, and rolled out to the broader team as the foundation for ongoing analytics.