Coincident index in Alaska

About this data

A bundle of nine Alaska indicators ingested through the single FRED API rather than five separate source pipelines: weekly initial unemployment claims (the fastest labor-market signal in the state), the Philadelphia Fed coincident and leading indexes (one-number summaries of the AK economy), FHFA house price indexes for Alaska and Anchorage, the Realtor.com Anchorage housing-market trio (median listing price, days on market, active listings — the best high-frequency AK housing signals anywhere), and seasonally adjusted statewide housing permits. FRED is the aggregation layer; each series has a different originating agency.

Known issues & gotchas

• One endpoint, N calls: GET https://api.stlouisfed.org/fred/series/observations?series_id={ID}&api_key={KEY}&file_type=json — one request per series, loop over the nine IDs. Rate limit ~120 requests/min, far above our needs. Missing values arrive as the string ".", not null — filter before float-parsing. All values are strings in JSON.
• Mixed frequencies in one dataset: weekly (claims), monthly (Realtor.com trio, permits, Philly Fed indexes), quarterly (both HPIs). The dataset-level freshness_max_gap (P120D) is set to tolerate the quarterly series; a per-series freshness check on initial_claims.ak.nsa (expect ≤ ~14-day gaps) is worth adding in the connector, since the dataset-level bound won't catch a stalled claims feed quickly.
• FRED's mirror of AKICLAIMS appeared to lag the DOL ETA source by weeks-to-months in a research snapshot (2026-08); verify actual update cadence at build time — if the lag is real, the DOL ETA report (oui.doleta.gov / dol.gov weekly claims release) is the fallback for the latest weeks.
• Realtor.com series get restated when their methodology changes (has happened before) — full-history refetch every run, don't append-only. Same for