Datasetpublic

stripeddonkey-data/nairobi-longitudinal-employment-career-mobility

Nairobi Longitudinal Employment & Career Mobility (emp-01) A 100% synthetic longitudinal employment dataset for a fixed cohort of 10,000 Nairobi city-twin citizens. It contains 13,947 lifecycle events covering formal-employment starts, promotions, terminations, and self-employment entry, with point-in-time education, professional-credential, and occupation-qualification context. Files File Rows Columns Grain data/employment_events.csv 13,947 48 One row… See the full description on the dataset page: https://huggingface.co/datasets/stripeddonkey-data/nairobi-longitudinal-employment-career-mobility.

sourceHugging Facecc-by-4.0updated 5d agoView on Hugging Face

0likes70downloads

Dataset Card

Nairobi Longitudinal Employment & Career Mobility (emp-01)

A 100% synthetic longitudinal employment dataset for a fixed cohort of 10,000 Nairobi city-twin citizens. It contains 13,947 lifecycle events covering formal-employment starts, promotions, terminations, and self-employment entry, with point-in-time education, professional-credential, and occupation-qualification context.

Files

File	Rows	Columns	Grain
`data/employment_events.csv`	13,947	48	One row per employment lifecycle event
`data/employment_person_summary.csv`	10,000	31	One row per citizen
`data/sample.csv`	25	31	Deterministic preview of the person-summary file

Event mix

Event type	Rows
`SELF_EMPLOYMENT_ENTRY`	5,677
`EMPLOYMENT_START`	5,161
`EMPLOYMENT_TERMINATION`	2,787
`PROMOTION`	322

7,368 citizens have exactly one event and 2,632 have two or more.

Education and occupation context

Education is evaluated as of each event date. Formal-employment rows can also carry an occupation-manifest crosswalk:

—OccupationCode
—OccupationMinimumQualification
—OccupationCertificationCode
—OccupationCertificationRule
—MeetsOccupationQualificationFloor

The qualification is a modeling floor, not the citizen's final education level and not a universal real-world legal rule.

At each citizen's latest published event:

Highest completed level	Citizens	Share
`SECONDARY`	8,543	85.43%
`BACHELOR`	876	8.76%
`CERTIFICATE`	230	2.30%
`DIPLOMA`	202	2.02%
`PRIMARY`	120	1.20%
Incomplete / NULL	29	0.29%

Among the 2,290 occupation-mapped citizens, 2,146 (93.7%) meet their modeled minimum qualification, 137 are below it, and 7 are unevaluable from the published evidence.

Logical splits

DatasetSplit is assigned once per citizen (Train, Validation, or Test). Every event for a citizen stays in the same logical split.

Validation

The shipped CSVs were revalidated directly:

—0 duplicate canonical events
—0 broken event sequences
—0 citizen split leakage
—0 events before date of birth
—0 AgeAtEvent mismatches
—0 formal terminations before episode start
—0 promotions outside their formal-employment episode
—0 overlapping formal-employment episode pairs
—0 future-dated education completions
—0 future-dated professional-credential issue dates
—0 person-summary/event reconciliation mismatches
—0 occupation-floor flag recomputation mismatches
—0 deterministic-preview mismatches

See `docs/CERTIFICATION.md`.

Use cases

—career-transition and next-event modeling
—time-to-promotion and time-to-termination analysis
—formal-employment vs. self-employment transition analysis
—point-in-time education and credential feature engineering
—occupation/qualification consistency analysis

Important notes

—Do not use identifiers such as CitizenId, SourceRecordId, ContractId, SelfEmploymentId, or EmploymentEpisodeId as predictive features.
—Order each citizen's history by EventSequence, not physical CSV row order.
—WardName is current synthetic residence, not historical residence reconstructed to each event.
—This dataset is synthetic and is not an estimate of real Nairobi employment or education rates.

Documentation

—`docs/DATA_DICTIONARY.md`
—`docs/METHODOLOGY.md`
—`docs/CERTIFICATION.md`
—`docs/KNOWN_LIMITATIONS.md`
—`docs/PROVENANCE.md`

License

CC BY 4.0. See `LICENSE` and `CITATION.cff`.