Groundfact — a governed semantic layer over official statistics

Year	Nepal (deaths per 1,000 live births)	India (deaths per 1,000 live births)	Bangladesh (deaths per 1,000 live births)	South Asia (deaths per 1,000 live births)	China (deaths per 1,000 live births)	Japan (deaths per 1,000 live births)
1990	138.4	127	146.5	128.6	53.6	6.3
1991	131.3	123.4	140.3	124.5	53	6.2
1992	124.3	119.9	134.1	120.6	52.1	6.1
1993	117.7	116.5	127.8	116.8	50.8	6
1994	111.4	113.1	121.6	113.1	49.2	5.9
1995	105.3	109.7	115.5	109.3	47.4	5.7
1996	99.5	106.2	109.4	105.5	45.5	5.5
1997	94	102.7	103.3	101.7	43.5	5.2
1998	88.7	99.1	97.1	97.8	41.4	5
1999	83.6	95.4	91.2	93.9	39.1	4.7
2000	78.7	91.8	85.5	90	36.6	4.5
2001	74.3	88.1	80.2	86.1	34.1	4.3
2002	70.2	84.6	75.3	82.4	31.4	4.1
2003	66.5	81.1	70.8	78.9	28.8	4
2004	63.1	77.7	66.8	75.6	26.3	3.9
2005	60	74.3	63	72	24	3.7
2006	57.3	71	59.6	68.7	21.9	3.6
2007	54.7	67.8	56.5	65.5	20.1	3.5
2008	52.4	64.5	53.6	62.4	18.5	3.4
2009	50.1	61.2	51.1	59.4	17.1	3.4
2010	47.9	58.1	48.7	56.3	15.7	3.3
2011	45.8	55	46.5	53.4	14.6	3.3
2012	43.7	52	44.4	50.5	13.5	3.1
2013	41.7	49.1	42.4	47.7	12.5	3
2014	39.7	46.4	40.5	45.1	11.6	2.9
2015	37.8	43.7	38.5	42.6	10.7	2.8
2016	36	41.3	36.7	40.3	10	2.7
2017	34.2	39	35	38.1	9.3	2.6
2018	32.6	36.9	33.5	36	8.6	2.5
2019	31	34.9	32.3	34.1	8	2.5
2020	29.6	33	31.4	32.4	7.5	2.4
2021	28.3	31.2	30.9	30.8	7	2.4
2022	27.1	29.5	30.7	29.3	6.6	2.4
2023	26.1	28	30.6	28	6.1	2.4
2024	25.1	26.6	30.5	26.8	5.7	2.4

What the figures say

In Nepal, under-five mortality fell from 138.4 in 1990 to 25.1 deaths per 1,000 live births in 2024 — an 81.9% fall. All 6 series shown fell over the same window: China furthest, by 89.4%, and Japan least, by 61.9%. Nepal sat above India until 1993 and has been below it in every year since 1994.

Computed from the rows above — first and last observation, change, ranking and crossings — not written by a model, so it cannot drift from the data.

The same country's numbers arrive from dozens of institutions in different shapes, on different schedules and under different licences. Groundfact consolidates them into a single queryable model — asked in plain English, with every figure traced back to its source, dataset and snapshot date. Nepal is the first instance; the architecture is country-agnostic.

See all of Nepal's data →

Ask it something

Every question below is answered from the exact-match cache this build already holds — no model call, no waiting, and no figure that was not already true before you arrived.

What is Nepal's under-five mortality rate?

27.1 deaths per 1,000 live births · 2022

World Bank · WDI · 2026-08-09

Answered from cache — no model call.

How much rice does Nepal produce?

5,955,500 tonnes · 2024

Nepal produced 5,955,500 tonnes of rice in 2024

FAOSTAT · QCL · 2026-08-09

Answered from cache — no model call.

What is Nepal's adult literacy rate?

68.7 % · 2019

68.7% of adults aged 15+ could read and write in 2019

World Bank · WDI · 2026-08-09

Answered from cache — no model call.

How much does Nepal spend on education?

3.69 % · 2024

Nepal spent 3.69% of GDP on education in 2024

World Bank · WDI · 2026-08-09

Answered from cache — no model call.

Ask your own question →

Where the data comes from

9 sources, each ingested on its own schedule under its own licence. The counts below are this build's, not a brochure's.

Sources in this build: what each one supplies, how far back it reaches, when Groundfact last refreshed it, and under what licence. A dash under Indicators means the source is loaded into the warehouse but not yet mapped into the semantic layer, so nothing on the site cites it yet.
Source	What we take	Indicators	Coverage	Last refreshed	Licence
World Bank data.worldbank.org	WDI refreshed quarterly	43 434,869 values	1960–2025 260 geographies	2026-08-09 Fresh	CC-BY-4.0
UN SDG Global Database unstats.un.org	211 datasets refreshed quarterly	2 13,660 values	1980–2026 1 geography	2026-08-08 Fresh	CC-BY-3.0-IGO
International Monetary Fund www.imf.org	5 datasets refreshed quarterly	5 44,477 values	1948–2031 202 geographies	2026-08-09 Fresh	attribution
FAOSTAT www.fao.org	QCL refreshed irregular	2 3,324 values	1961–2024 26 geographies	2026-08-09 Fresh	CC-BY-4.0
ILOSTAT ilostat.ilo.org	10 datasets refreshed weekly	3 64,874 values	2000–2025 12 geographies	2026-08-08 Fresh	CC-BY-4.0
UNESCO Institute for Statistics uis.unesco.org	UIS refreshed quarterly	7 24,196 values	1970–2025 210 geographies	2026-08-09 Fresh	CC-BY-SA-4.0
DHS Program dhsprogram.com	87 datasets refreshed irregular	3 1,059 values	1985–2025 87 geographies	2026-08-09 Fresh	DHS-aggregate-attribution
Our World in Data ourworldindata.org	9 datasets refreshed monthly	8 2,640 values	1750–2025 207 geographies	2026-08-09 Fresh	CC-BY-4.0
Humanitarian Data Exchange data.humdata.org	4 datasets refreshed monthly	— 57,384 values loaded	2001–2026 191 geographies	2026-08-09 Fresh	per-dataset

Where sources disagree

When official sources disagree, Groundfact doesn't quietly pick one. It shows both — and the honest reason they differ. Neither source is wrong.

Reconciliation spotlight Two sources, two numbers

Unemployment, total (% of total labor force), NPL, 2008

1.38–10.6 % two sources differ by 9.22 percentage points

Why do these differ?

ILOSTAT's UNE_DEAP_SEX_AGE_RT is the direct labour force survey estimate (total, ages 15+), reported only in the years Nepal ran an NLFS round (1996, 1999, 2003, 2008, 2015, 2017); the World Bank's SL.UEM.TOTL.ZS is a smoothed modelled annual series. The two converge almost exactly in 2017 (verified live: 10.66% both) but ILOSTAT sits far below the Bank's figure before it (1.4-4.5%). That is a real break in Nepal's own survey methodology, not a data error: Nepal's LFS moved from an extended activity definition (which classified most subsistence/unpaid economic activity as employment, pushing measured unemployment very low) to the standard ICLS-19 definition around the 2017/18 NLFS III round.

Lower reading

1.38

ILOSTAT

ILOSTAT · UNE_DEAP_SEX_AGE_RT_A · 2026-08-08

Higher reading

10.6

World Bank

World Bank · WDI · 2026-08-09

Read the reconciliation note →

How it works

Sources

Official statistics — World Bank, UN agencies, national statistics offices and more — ingested on a schedule, each one licensed and attributed.
Semantic layer

A curated model reconciles indicator definitions, units and geographies across sources, so "under-five mortality" means one thing everywhere it appears.
Cited answers

Every figure carries its source, dataset and snapshot date — inline, not in a footnote you have to go looking for.

Read the architecture →

Nepal's under-five mortality has fallen 82% since 1990.

Live from this build

Under-five mortality

What the figures say

Ask it something

Where the data comes from

Where sources disagree

Unemployment, total (% of total labor force), NPL, 2008

How it works

Sources

Semantic layer

Cited answers