Interactive evidence intelligence platform prototype built on Afrobarometer Round 10 Ghana survey microdata
# Afrobarometer Insight Platform — 39-Country Round 9 Build
An interactive evidence intelligence platform built on Afrobarometer's real
**Round 9 Merged Data** — 53,444 interviews across all 39 fielded countries.
Developed independently in connection with a proposal for Afrobarometer's
2026 Flagship Report interactive dashboard.
**Live demo:**
fadl462.github.io
**Live locally:** open `index.html` in any modern browser, or serve the
repo root with any static file server (no build step, no server-side code —
everything runs client-side):
```bash
python3 -m http.server 8000
# then open
localhost
```
## What this is
The platform ships two tiers of data:
1. **Precomputed cross-country aggregates** (`data/country_aggregates.json`,
`data/continental.json`) — small, fast-loading files covering every
country on every indicator, so the Executive Briefing and Compare
Countries views render instantly with no per-country fetch.
2. **Full respondent-level microdata, per country** (`data/countries/*.json`)
— a columnar, integer-coded file for each of the 39 countries (~700–900KB
raw, ~100KB gzipped), fetched on demand whenever you change the sidebar's
**focus country** selector. This powers the deep-dive views (Indicator
Explorer, Demographic Intelligence, Regional Intelligence, Policy Insight
Centre, and within-country Ask-the-Data questions) with genuine
respondent-level computation, not pre-baked charts.
Every statistic is traceable: single-country figures use `withinwt_hh`
(Afrobarometer's within-country weight); cross-country and continental
figures use `Combinwt_new_hh` (Afrobarometer's own multi-country weighting
factor, which gives each of the 39 countries equal say in the continental
average regardless of population — matching Afrobarometer's published
cross-country reporting convention).
### Modules
- **Executive Briefing** — continental headline KPIs, a rule-based insights
feed, the top problems citizens name …