Methodology
How the prototype classifies, matches, and cites data
This page documents the public datasets, confidence labels, keyword matching, aggregation rules, and known limitations behind the headline metrics.
Prototype data refresh: May 14, 2026
Confidence labels
Use these labels to separate agency-published numbers from computed or matched proxies.
Published verbatim by a named public agency.
Computed from official public inputs, such as summing WIOA streams.
Joined or filtered across public datasets; useful proxy, not a definitive ledger.
Partial or conservatively extracted source context that needs human review before citation.
Not centrally public in the sources reviewed.
Source inventory and data vintage
Each major page links back to these public sources near the metric it supports.
| Dataset | Use in prototype | Vintage / cadence | Source |
|---|---|---|---|
| NYSDOL LWDA allocations | Official PY2025 Adult, Dislocated Worker, Youth, and combined formula allocations by LWDA. | Annual PY2025 allocation PDF; source notes TEGL official allocations dated May 20, 2025 | NYSDOL allocation PDF |
| NYSDOL performance reports | Parsed PY2025 Q2 LWDA-level employment Q2/Q4 and median earnings metrics. | Quarterly PY2025 Q2 report covering July 2025 through December 2025 | NYSDOL performance reports |
| TrainingProviderResults.gov | NY eligible training providers/programs, WIOA ITA counts, outcomes, and cost fields. | Ongoing API refresh Fetched from the public TrainingProviderResults.gov API during the latest prototype refresh | TrainingProviderResults.gov |
| NYC DYCD contracts | Contract fiscal-year rows filtered by WIOA/workforce/youth terms as a public contract proxy. | NYC Open Data refresh Fetched from NYC Open Data during the latest prototype refresh | NYC DYCD contracts |
| NYC DYCD program sites | Geocoded Train & Earn, Learn & Earn, and WIOA youth service sites. | NYC Open Data refresh Fetched from NYC Open Data during the latest prototype refresh | NYC DYCD sites |
| NYSDOL career centers | Career center locations, contacts, and workforce regions. | data.ny.gov refresh Fetched from data.ny.gov during the latest prototype refresh | NYSDOL career centers |
NYC DYCD contract proxy method
MatchedDYCD contract rows are filtered when the solicitation name, solicitation detail, or major program contains one of these terms:
The resulting fiscal-year rows are summed as a public contract proxy. This is not a one-to-one WIOA expenditure ledger: contracts can span years, include adjacent workforce programs, or omit locally held WIOA obligations.
ETP matching and suppression notes
Training provider records are pulled from the public USDOL API for `field_state = NY`, sorted by stable ID, and de-duplicated by API record ID. Providers/programs are mapped to the nearest LWDA centroid from their published coordinates.
`N/A` in provider tables means the public API supplied a blank, negative sentinel, zero/one-dollar placeholder, suppressed, or otherwise non-actionable value. It should not be read as proven zero cost or zero performance.
Known unresolved questions
- Statewide provider payment ledgers outside NYC are not centrally public in the reviewed sources.
- Local contract/subrecipient documents vary by board and are not normalized statewide.
- Participant-level or de-identified cohort files are needed to connect service activity to outcomes precisely.
- ETP suppression rules and low-cohort reporting can make provider performance fields unavailable.