Open Data14 Stars

Open Data Uganda

Standardized open data clearinghouse for Ugandan demographics & public health

Centralized public data democratization hub standardizing 14+ public datasets from UBOS, UN OCHA, WHO, and the World Bank. Distributed in clean, machine-readable open formats (.csv and .jsonl) with automated GitHub Actions dataset validation and Python Pandas loading scripts.

Visit Live Site View Source

CategoryOpen Data

LinksLive + Source

StackData Engineering, JavaScript

Scale14 Available

The Challenge

Crucial public datasets in Uganda across district populations, administrative boundaries, health facilities, and economic indicators were fragmented in locked PDF reports, obsolete portal endpoints, or inconsistent spreadsheets, making data science and civic tech development prohibitively slow.

Architecture & Solution

Built a central open-access repository curating, normalizing, and versioning verified data from UBOS, UN OCHA, WFP, and World Bank. Published standardized `.csv` and `.jsonl` schemas backed by continuous CI validation workflows that check row schemas, geographic codes, and null-safety.

System Architecture

┌────────────────────────────────────────────────────────┐
│                   Open Data Uganda                     │
├────────────────────────────────────────────────────────┤
│ Sources: UBOS · UN OCHA · World Bank · WHO · UNESCO    │
└──────────────────────────┬─────────────────────────────┘
                           │
             Automated Cleaning & Parsing
                           │
             ┌─────────────┴─────────────┐
             │ GitHub Actions Validation │
             │ • Schema integrity checks │
             │ • Geo-boundary verification│
             │ • Encoding & null audits  │
             └─────────────┬─────────────┘
                           │
      ┌────────────────────┴────────────────────┐
      ▼                                         ▼
Standardized CSV                         Streamable JSONL
(Pandas direct URLs)                     (High-throughput ingest)

Engineering Decisions

Zero-Dependency Direct URL LoadingStructured repository releases and raw URLs so data scientists can ingest clean datasets directly into Python Pandas or R with a single `pd.read_csv(url)` call without API keys.
Dual Format DistributionSupplied both tabular `.csv` for business analysis and newline-delimited `.jsonl` for large-scale data pipelines and machine learning training corpora.
Automated Schema Linting on PRsConfigured CI actions to validate every incoming dataset contribution against strict column definitions and geographic identifier standards.

Outcomes

14+ standardized public datasets actively used by civic tech developers and researchers
14 organic stars on GitHub with open contribution guidelines under CC BY 4.0 license
Saves researchers hours of manual data extraction from government PDF bulletins