11 · Personal & Academic Project
HIV Epidemiological Database
Creating a governed foundation for cross-country analysis
Business value
Researchers could perform reproducible longitudinal and cross-country analysis from one documented database.
01 · Business problem
CDC, World Bank, and UNAIDS published data in different formats, country conventions, indicator definitions, and time ranges.
02 · My approach
I treated source integration as a governance problem, resolving entities, indicators, lineage, and reproducibility.
03 · What I built
Five integrated sources, six normalized MySQL/PostgreSQL tables, ER schemas, data dictionaries, and governance rules.
What changed my thinking
The corrections, tradeoffsand decisions behind it.
- 01
Country names were data-quality issues
Small naming differences prevented reliable joins across global sources.
- 02
Definitions mattered as much as values
Indicators with similar names did not always measure the same thing.
- 03
Documentation was part of the product
Lineage and data dictionaries made the database reusable by others.