Support this project

SparkyData is built and maintained by one person, in the open, and given away. If it has been useful to you, a contribution keeps the next release coming.

Every published statistic is a constraint, and the collection of constraints is never finished: each new survey release, each new source, each new geography is another rebuild — refetching sources, recalibrating, re-validating, and writing up what changed. Contributions buy that time.

SparkyData is not a charity. It is a for-profit sole proprietorship. A contribution is a voluntary, one-time tip for work you already have: it is not tax-deductible and does not entitle you to any product, service, or access. Payments are handled by Stripe; we never see your card details.

What is in store next

• The first calibrated release, v2026.1. A full New York metro population built from the 2020–2024 American Community Survey microdata, with its validation report and source manifest.
• A larger population for people running simulations rather than tests — the same pipeline at a hundred thousand households and beyond.
• More of the balance sheet. Net worth and debt by type, health and insurance, detailed commute and workplace.
• The time axis. A multi-decade annual history in which people arrive, age, earn, move and leave at recorded rates — the part that lets a cohort be followed rather than a cross-section compared.
• A second metro. Once the method is settled on New York, the same pipeline pointed at another region's microdata and published statistics.
• More articles, each with its dataset and code, on the questions the published tables cannot answer on their own.

The roadmap, stated up front

Every release becomes free under CC BY 4.0 twelve months after it is published. Eventually the newest release will be a paid product — that is how the work stays funded. Two commitments go with that: nothing already published free will ever become paid, and the twelve-month clock is not restarted by minor releases. Saying this now rather than later is deliberate. A paywall that appears without warning is a betrayal; one that was on the roadmap from the first day is a business model.

Ways to help that cost nothing

• Cite the dataset if you publish something built on it.
• Tell us where the data is wrong — especially against the validation report — at sparkytech.dev@gmail.com.
• Tell us what variable you needed and could not find.