PROJECT 05 / KENYAN PUBLIC DATA · CIVIC UNDERSTANDING
A vote. A budget. A bill.
What do the numbers actually tell us?
One headline can hide several different questions. Follow the vote, the budget and the bill counts back to their definitions.
2022 election · FY 2023/24 records · Separate periodsNO POLITICAL BACKGROUND REQUIRED
You hear that turnout was 64.77%, counties spent billions, or Parliament processed 128 bills. Before deciding what any of that means, ask: who was counted, out of what total, and at which stage?
I use public records to explain participation, spending and parliamentary activity, keeping each measure’s denominator and scope visible.
How does the meaning of a public headline change when we reveal its denominator, scope and stage?
This is a deliberately fixed historical study, not a live tracker or a statement of current office-holders’ performance. Election and financial-year records describe different periods and are not treated as a causal chain.
01 / THE VOTE · 2022
64.77%
Declared valid votes plus rejected presidential ballots, divided by the declared national registered electorate.
14,326,751 ballots / 22,120,458 registered voters. The difference is 7,793,707 registered voters not recorded voting in these totals. It is not a measure of why people did not vote.
Population: includes people who are not eligible or registered to vote. It is not this denominator.
Registered electorate: the people listed in the election register.
Ballots recorded: valid votes plus rejected ballots for this particular election.
A county’s turnout does not establish how its young people, women or any ethnic group voted. These are aggregate records.
Source: IEBC declaration dated 15 August 2022, reproduced in Carter Center Appendix A, PDF pages 50–52. Diaspora is included in the national total and kept separate from the 47-county explorer. Published arithmetic differences are disclosed below.
Kenya’s National Assembly has 290 constituency seats, 47 county women’s seats and 12 nominated seats, plus the Speaker. The Senate has 47 elected county seats and 20 nominated seats, plus the Speaker. These are different routes to representation; presidential turnout does not tell us who holds those seats.
Structure explained by Parliament of Kenya. Seat categories are not a count of currently occupied seats or a party-strength estimate.
LOOK CLOSER / YOUR COUNTY, WITH CONTEXT
The filled portion shows ballots recorded or budget spent. The pale remainder shows the rest of that denominator. Both charts run from 0 to 100%. The two periods are shown together for convenience, not to imply turnout caused budget outcomes.
02 / THE MONEY · JULY 2023 TO JUNE 2024
The first two are in the financial table. The third needs additional evidence.
57.5%
KES 109.23 billion spent against KES 189.93 billion approved across the 47 county governments.
24.4%
The same KES 109.23 billion, now divided by KES 446.76 billion in total expenditure. Same numerator. Different question.
Which projects were delivered, to what standard, and for whom? Check implementation records, completion evidence and audit findings before making a service-delivery claim.
“Absorption” means spending as a proportion of the approved budget. A larger rate alone does not demonstrate better services. A budget-to-spending gap is not, by itself, evidence of missing money. Funding availability, timing, procurement and implementation require separate investigation.
Source: Controller of Budget, FY 2023/24, Table 2.3, printed pages 6–7, accessed through an Isiolo Municipality-hosted copy. Calculations use the table’s KES million values. This is reported expenditure, not an independent audit by this project.
03 / THE WORK · FY 2023/24
The National Assembly reported 128 bills processed. The report describes bills introduced and considered at various stages. Calling that “128 laws passed” would change the claim.
| Activity | National Assembly | Senate |
|---|---|---|
| Bills processed | 128 | 60 |
| Questions processed | 525 | 173 |
| Public petitions processed | 44 | 21 |
Do not add the Houses’ counts to claim unique bills or petitions. The same item can move between Houses. Different mandates and processing definitions also prevent these counts from serving as a productivity league table.
For a named bill, trace its stage and any assent record. For a question or petition, find the response, committee report or recorded resolution. Activity counts alone cannot show whose concerns were resolved.
Source: Parliamentary Service Commission annual report FY 2023/24, Tables 5.2 and 5.10, printed pages 69 and 77 (PDF pages 81 and 89). Senate sitting totals conflict within the report and are excluded.
A REPORTING STANDARD
The recommendation is a reporting standard: every public statistic should show its period, unit, denominator or processing stage, source, and limit. That is a practical requirement a civic-information team could test.
The case connects requirements thinking, data reconciliation and clear communication. Its proposed reporting standard is ready for a reader-comprehension test.
EVIDENCE / METHODS / LIMITS
The charts have a paper trail. Here’s the analysis, the data and the source record behind this case.
Development budget spent · FY 2023/24
Three-county extract · 0–100% scale
Controller of Budget, Table 2.3
| County | Approved | Spent |
|---|---|---|
| Baringo | 3,286.70 | 2,115.95 |
| Bomet | 2,386.79 | 1,819.86 |
| Bungoma | 5,284.31 | 3,113.59 |
First three records from the included county-budget CSV. Full file includes 47 counties and the published total.
Different records, periods and denominators. The methods explain the boundaries.
The reproduced election declaration’s geographic registered-voter rows sum to 22,119,458, while its national row states 22,120,458, a difference of 1,000. National turnout uses the stated national row; county charts use their own stated rows. I preserve the difference rather than allocating it to a county.
In Kilifi and West Pokot, published valid-vote totals exceed the sum of the four candidate columns by 10 and 100 respectively. The explorer preserves published valid totals, flags the affected counties and does not use this discrepancy to infer intent, fraud or electoral impact.
The budget narrative reports overall absorption as 79.5%, while Table 2.3 gives 79.4%. Recalculation from that table rounds to 79.4%. Budget entries also contain ordinary rounding differences. The Senate sitting narrative implies 124 sittings while its table gives 120, so sittings are not used.
The reviewed work includes ELOG’s election comparisons, IBP Kenya’s budget transparency survey and Samy Migwi’s voter dashboard repository. Their questions include election patterns, document transparency and geographic vote contribution.
This case connects those reading skills: follow a headline through its denominator, administrative stage and evidence limits.
Election CSV totals are credited to Paul-Oti’s Kenya Election Data, pinned to the commit in the source manifest. Its MIT licence is retained. Registered counts were transcribed from the IEBC declaration reproduced by the Carter Center, rather than mixed with the pre-election county register.
Budget figures were extracted from Controller of Budget Table 2.3. Parliamentary figures were transcribed from the PSC report. Original PDFs are linked, not republished. Source files were retrieved on 11 October 2026. The study deliberately uses historical periods and makes no claim to use the latest available data.
Presidential turnout uses valid plus rejected ballots divided by the county’s declared registered voters. Development absorption uses spending divided by approved development budget. Different periods; no causal comparison.
| County | 2022 presidential turnout | FY 2023/24 development absorption |
|---|---|---|
| Mombasa | 43.76% | 43.4% |
| Kwale | 54.94% | 48.3% |
| Kilifi | 49.03% | 55.8% |
| Tana River | 67.02% | 57.8% |
| Lamu | 62.55% | 52.2% |
| Taita Taveta | 61.77% | 46.9% |
| Garissa | 54.86% | 44.8% |
| Wajir | 64.69% | 83.4% |
| Mandera | 62.85% | 82.6% |
| Marsabit | 69.12% | 68.3% |
| Isiolo | 66.47% | 55.5% |
| Meru | 66.18% | 79.3% |
| Tharaka-Nithi | 70.08% | 54.5% |
| Embu | 66.61% | 60.3% |
| Kitui | 62.38% | 71.3% |
| Machakos | 60.20% | 58.6% |
| Makueni | 61.05% | 62.2% |
| Nyandarua | 67.06% | 45.9% |
| Nyeri | 68.29% | 72.4% |
| Kirinyaga | 69.85% | 53.5% |
| Murang'a | 68.10% | 65.6% |
| Kiambu | 65.15% | 49.5% |
| Turkana | 60.63% | 63.1% |
| West Pokot | 79.51% | 67.3% |
| Samburu | 70.65% | 60.8% |
| Trans Nzoia | 63.37% | 59.8% |
| Uasin Gishu | 69.51% | 48.9% |
| Elgeyo-Marakwet | 77.96% | 49.3% |
| Nandi | 76.05% | 49.8% |
| Baringo | 77.59% | 64.4% |
| Laikipia | 64.73% | 48.7% |
| Nakuru | 65.53% | 46.0% |
| Narok | 77.73% | 90.3% |
| Kajiado | 66.96% | 57.5% |
| Kericho | 78.56% | 61.7% |
| Bomet | 79.88% | 76.2% |
| Kakamega | 60.29% | 57.3% |
| Vihiga | 60.13% | 50.9% |
| Bungoma | 63.51% | 58.9% |
| Busia | 67.08% | 59.6% |
| Siaya | 70.89% | 64.6% |
| Kisumu | 71.37% | 37.0% |
| Homa Bay | 73.70% | 86.3% |
| Migori | 74.49% | 47.3% |
| Kisii | 63.92% | 29.2% |
| Nyamira | 65.16% | 61.2% |
| Nairobi City | 55.98% | 38.7% |
THE TAKEAWAY
A useful political analysis gives you the context to question a claim, and the evidence to check the answer.