Enterprise Data Transformation and Analytics Platform

Pakuwon Group, one of Indonesia's largest real estate and mall operators, runs a nationwide portfolio of malls, offices, and high-rise residences with annual maintenance spend north of IDR 5 billion. Management wanted to run the portfolio on data rather than instinct, and to lay the groundwork for AI, but the numbers lived in disconnected systems and manual spreadsheets.

It was the group's first formal data transformation, and the starting point was tough:

• Operational and maintenance data fragmented across departments and individual buildings
• Records kept in siloed, manual, tabular formats with little analytical value
• The same asset named differently across buildings and files, so no cost could be summed across the portfolio
• No centralised data architecture or acquisition pipelines
• Tight timelines set by management's transformation agenda
• Almost no visibility into asset performance, maintenance efficiency, or cost drivers

VentureSEA acted as the end-to-end data transformation partner and delivered in disciplined phases, with the first working dashboards in front of management early rather than after a long build.

• Ran an as-is assessment and gap analysis, then defined a target-state architecture aligned to business goals
• Designed a centralised lakehouse on AWS, with a Redshift warehouse, purpose-built data marts, access controls, and data lineage from source to dashboard
• Built ingestion and ETL with AWS Glue to consolidate and standardise data from multiple source systems in near real time, with schema enforcement and quality checks at each stage so bad records are caught at the boundary, not discovered in a board pack
• Resolved asset identity across buildings and source files into a single asset dimension, the unglamorous entity-matching work that makes portfolio-level cost analysis possible at all
• Backfilled years of historical spreadsheet records into the same model, so trend analysis started with history on day one instead of waiting to accumulate it
• Delivered interactive QuickSight dashboards tailored to management, operations, and maintenance teams
• Instrumented the pipelines with freshness and quality monitoring, so a stalled source surfaces as an alert rather than as a stale number in a meeting
• Stood up a data governance working group with the group's IT and operations leads, and trained internal analysts to own dashboards and data quality after handover