knbs ke.knbs.energy-balance.annual annual

Energy Balance by Primary Source.

↗ source ↓ download

VERSION v2026-08-18

UPDATED 19d ago

CHECKED 1d ago

COVERAGE 2021 - 2024

SHAPE 84 × 9

SIZE 7.4 KB

FREQUENCY annual

PIPELINE automated

LICENSE Not stated

SOURCE www.knbs.or.ke

Kenya's whole energy system in one table, terajoules: how much coal, petroleum, electricity and biomass the country produced, imported, exported and consumed, for the four years to 2024, the shortest span on this seam. 84 observations across 21 items: 18 named sources, the total demand row, and two memorandum rows that are not terajoules at all.

Biomass dominates. Non-Renewable Feedstocks is 94.4 per cent of total demand in 2024, 1,341,522.4 of 1,421,441.4 terajoules, and Fuelwood alone, one component of that feedstocks figure, is 84.1 per cent on its own.

Non-Renewable Feedstocks more than doubles between 2022 and 2023, from 562,953.3 to 1,303,872.7 terajoules, 2.32 times, and Per Capita Consumption In Terms Of Giga Joules (Gj) roughly doubles with it, 12.5 to 26.2. Read as a trend line this looks like Kenyans burning twice the wood in one year; it is not. A change of that size in one year, in a series that moves gradually in every other year on file, is the shape of a methodology change, not a change in what households do. Neither Table 9.8's own footnotes, nor the workbook's own cover page, nor the published chapter's text, so far as this pipeline could reach it given the same certificate issue pipeline/lib/tls.py exists to work around, names a cause for this break or confirms that a new survey round landed in 2023. What the workbook's cover page does state, under Data Sources, is that Biomass energy demand specifically, unlike every other row in this table, is estimated from consumption patterns drawn from KNBS's own censuses and surveys rather than extracted from administrative data. That is why this series, and not petroleum or electricity, is the one capable of moving in a one-year step of this shape: an updated survey round can move a consumption estimate at once, where administrative records accumulate gradually. It is not confirmation that a survey update is what happened here, and this description does not claim it is, only that this series is structurally capable of the kind of step every other row in the table is not.

The total row states its own derivation: TOTAL DEMAND=(a+b-c+d+e+f), printed in the row's own label, with the matching tag printed on six other rows, Coal and Coke through Non-Renewable Feedstocks. That formula is asserted rather than assumed, parsed from the row labels themselves, in every year: 2024 sums to 1,421,441.4 terajoules exactly as printed. This is the only table on the whole seam that states its own derivation this way, and the only prior edition that does not: the 2025-02 edition's total row reads only "TOTAL Demand", no formula, no tags anywhere in the table, so this identity can only be checked on the current edition. A second identity, unrelated to the formula and holding on every edition on file, is that the six local-production rows (hydro, geothermal, thermal, wind, solar, co-generation) sum to Total Local Energy Production.

source_level mixes component, subtotal, total and memorandum rows in one column, and section mixes terajoules with a percentage and a per-capita gigajoule figure: summing across either is meaningless. Filter to one section, and within it to one source_level, before summing anything. The two memorandum rows, Local Production As Percentage of Total and Per Capita Consumption In Terms Of Giga Joules (Gj), are kept rather than dropped precisely because neither is reproducible from any row printed above it.

Two rows, Imports of crude oil and Stock changes and balancing item, print a real 0 in every year of the current edition but "-" in every year of the previous one: a genuine difference between editions, not a parsing gap.

LOADING DATA...

WHAT KENYA BURNS

Biomass is 94 per cent of the energy Kenya uses, and fuelwood alone is 84 per cent of it.

Energy demand by primary source · terajoules · 2021 to 2024

SOURCE: Kenya National Bureau of Statistics · Energy Balance by Primary Source (annual) · Updated 2026-08-18

#	year	period_start	section	source	source_path	source_level	status	value	unit
loading...

COLUMN	TYPE	DESCRIPTION	UNIT
year	int64	Calendar year	-
period_start	date	1 January of the year	-
section	string	The primary-source grouping KNBS's own banners organise the table by: coal_and_coke, liquid_fuels, electrical_energy, biomass, total_demand, or memorandum. Sections mix units for memorandum only (percent and gigajoule_per_person beside terajoule everywhere else); summing across section is meaningless regardless.	-
source	string	The row's own name. Carries the current edition's own printed tag, "(a)" through "(f)", on the six rows the total-demand formula names; the previous edition prints none of these tags at all, and this column reflects that difference rather than hiding it.	-
source_path	string	"{section} > {source}" with any tag suffix removed. The key this dataset is unique on, alongside year, and stable across editions even where a tag is not.	-
source_level	string	component, subtotal (Total Local Energy Production and Non-Renewable Feedstocks, each the sum of named rows beneath it), total (the row stating its own formula), or memorandum. Mixes levels that must not be summed together; filter to one before summing anything.	-
status	string	final or provisional	-
value	float64	The figure, in the row's own unit column. Null where KNBS prints no figure ("-" or ".."); a printed 0 is a real zero, not a gap.	-
unit	string	terajoule for every section except memorandum, where it is percent (Local Production As Percentage of Total) or gigajoule_per_person (Per Capita Consumption In Terms Of Giga Joules (Gj)).	-

CITATION · APA 7

imibare contributors. (2026). Energy Balance by Primary Source [Data set]. imibare. v2026-08-18. https://imibare.org/datasets/ke.knbs.energy-balance.annual