How the record is assembled.

Two CT Office of Early Childhood datasets on data.ct.gov: the licensing file (provider identity, capacity, status) and the investigations log (case-level records, including resolution and related inspection actions). Every field on a provider page maps back to one of these.

A daily job normalizes the feeds into a single provider record — keyed on the OEC credential identifier — and flags cases whose case number appears in more than one feed.

We render everything the state publishes: case numbers, dates opened and closed, how the case was received, resolution text, and inspection actions when the state links them to the same case ID. We do not paraphrase, summarize, or withhold detail from the state's record.

What isn't on these public feeds — full complaint narratives, inspector field notes, photographic documentation — requires a Freedom of Information request. Where a record is likely to have more behind it, we mark it FOI candidate and link to the FOI Desk.

Counts shown on a provider page are literal counts from the record. Comparisons to a license-type average and statewide average are computed across every active provider; they are context, not scores.

We deliberately do not publish a rating, letter grade, star count, or ranked list. Investigation counts alone cannot tell you whether a provider is good or bad. The record tells you what's been formally reported and resolved; your judgement does the rest.

If a provider believes a record is wrong, the correction path is with OEC — we reproduce the state's record; we don't maintain it. We reflect the current state feed within 24 hours of OEC updating it.

If you think a page on our site is misrepresenting the state's record (rendering error, miscategorized field, broken link), email the corrections desk.

Family child care homes operate from residential addresses. To protect provider privacy, we show only the city and zip code by default. Full addresses are available on request behind a click. Child care centers, group homes, and camps show full addresses as they are commercial locations.

Every field we render is sourced from data.ct.gov — the same feed that powers eLicense and 211 Child Care. If we show something that contradicts OEC's own systems, the bug is ours to fix. Here's how to spot-check any record on this site:

1. License status. Note the license number on any provider page. Search by license number at eLicense CT. Active/inactive, effective dates, capacity should match exactly.
2. Inspection history. For the same provider, open 211 Child Care and click Inspection History. Visit dates and finding codes there should appear on our provider page.
3. Statute language. Every finding we show is stamped with the CT regulation code in brackets (e.g. [19a-79-3a(b)(8)(B)]). Look it up in OEC's regulations PDF. The text we display is verbatim from the source feed.
4. Sync time. The footer of every provider page shows when we last pulled the record. Our daily job refreshes the full dataset around 06:00 UTC.

A finding (our term: N findings on this case) is a line-item on OEC's compliance checklist that a visiting inspector marked out-of-compliance. It is not a parent complaint. Most findings are paperwork (staff file missing a signed form), routine maintenance (fire extinguisher tag past due), or ratio/configuration items from a checklist that runs 130+ lines per visit.

A long-running center accumulates dozens of findings across routine inspections over years. That's expected — and is not the same signal as repeated complaint investigations, CAP referrals, or legal-division referrals, which we count separately in the benchmark box on each provider page.

On each finding we render: the statute code, the full statute language, the state's description code, the visit date, and a category label (Paperwork, Facility, Health & safety, Ratio). The category labels are ours; everything else is verbatim from the state record.

We chose Connecticut because OEC publishes cleanly, because the state's child-care licensing is concentrated in one office, and because the dataset is the right size to get right. Each additional state is its own integration job: different schemas, different disclosure laws, different case taxonomies. We'll announce each state as it goes live.