Data audit
| file | SHA-256 | rows (excl. header) | cols |
|---|---|---|---|
| data/wb_indicators.csv | 6823f02ed4c20fd12a54b8ca4454f85ba333f22337e2a9fc57c3cabccfdbfcd0 | 2349 | 5 (iso3,name,year,gdp_pc_ppp,life_expectancy) |
| data/wb_entities.csv | 06af6aec15122fecea5be5eb21464c505188ecc135bc55104a40f20f65040f8d | 296 | 5 (iso3,name,region,income_level,lending_type) |
| data/README.md | 76cb5af488d97339fb6319a5c9a7ba6ec98112ab6a0a8116e7bec6acd1f9f67a | – | – |
| Hashes match data/README.md and bundle.json. |
- Years 2015–2023; 261 entities in indicators, of which 44 are Aggregates (396 rows) → 217 countries/economies, each with all 9 years (1953 rows).
- No duplicate (iso3, year) rows; no duplicate iso3 in entities; every indicator iso3 exists in entities (35 entity-only aggregate codes absent from indicators — harmless).
- Units: GDPpc PPP constant 2021 intl $ (range ~992 to ~139,541, all >0); LE in years.
- Missing: gdp_pc_ppp 164 missing (all entities); life_expectancy 0 missing. Countries with both indicators: 199 each year 2015–2022, 197 in 2023 → analysis year 2023 (>=150).
- 2023 missing GDP (20): American Samoa, British Virgin Islands, Channel Islands, Cuba, Eritrea, French Polynesia, Gibraltar, Guam, Isle of Man, DPRK, Liechtenstein, Monaco, New Caledonia, N. Mariana Islands, South Sudan, St. Martin, Syria, Venezuela, US Virgin Islands, Yemen.
- Anomalies: CAF life expectancy series erratic (2019 31.5, 2022 18.8, 2023 57.4); PSE 2023 drop of 11.5 years (Gaza war); SSD 2015–2018 swings (SSD not in 2023 sample). CAF and PSE 2023 are in the sample; kept as reported, sensitivity run excluding them.
- Parsing: read with keep_default_na=False so only empty cells are missing.
- Information-availability: no input violates the rules; aggregates excluded per rule. No imputation.