090°Données ouvertes
Jeux de données ouverts, entièrement documentés — interrogeables ici, et lisibles par n’importe quel LLM.
Les titres et les descriptions proviennent des sources de données, en anglais.
4 jeux de données
U.S. Bureau of Economic Analysis
US metro price-level intelligence from the BEA Regional Price Parities (RPP, 2008–2024 vintage, keyless FRED redistribution, U.S. public domain): ~384 metropolitan statistical areas (OMB 2023 delineations), annual, US = 100. Method: the closed RPPALL<CBSA> universe is pulled via fredgraph.csv; every series must span 2008 to the latest vintage with zero gaps and plausible values or the ingest fails loudly. Derived signals per metro-year: cost_premium_pct (RPP − 100), a documented 5-band price_tier (very_low <90 … very_high ≥110), within-year national_rank (1 = most expensive), 1-year and 5-year index-point drift (rpp_yoy_pp / rpp_5y_pp), and local_purchasing_power_100 (10000/RPP: the national-dollar value of $100 spent locally). Use as prediction features: join to customers/orders/leads by cbsa_code + year for territory-level demand models, site selection, regional pricing, and real (price-level-adjusted) market sizing — the state panel's coarser sibling for metro granularity. Caveats: RPPs compare price levels across places, not inflation over time; the latest 1–2 vintages are routinely revised.
US state price-level intelligence from the BEA Regional Price Parities (RPP, 2008–2024 vintage, keyless FRED redistribution, U.S. public domain): 50 states plus DC, annual, US = 100. Method: the closed 51-series <ST>RPPALL universe is pulled via fredgraph.csv; every series must span 2008 to the latest vintage with zero gaps and plausible values or the ingest fails loudly. Derived signals per state-year: cost_premium_pct (RPP − 100), a documented 5-band price_tier (very_low <90 … very_high ≥110), within-year national_rank (1 = most expensive), 1-year and 5-year index-point drift (rpp_yoy_pp / rpp_5y_pp — relative price convergence or divergence), and local_purchasing_power_100 (10000/RPP: the national-dollar value of $100 spent locally). Use as prediction features: join to customers/orders/leads by state_code + year to deflate nominal amounts into real terms, set regional price lists and sales quotas, and score market affordability — a $29 subscription feels like $26 in California but $33 in Mississippi. Caveats: RPPs compare price levels across places, not inflation over time; the latest 1–2 vintages are routinely revised.
State GDP Signals (derived)
Annual US state GDP signals from FRED/BEA (1997 -> latest): all-industry total GDP for the 50 states + DC with 1-year and 10-year growth rates, 10-year growth volatility, 3-sigma anomaly flags, drift forecasts, 10-year growth z-scores and gauges: each state's share of US GDP, a contraction flag and an outperform-vs-US flag. The output lens on regional growth — the GDP companion to the coincident index in fred-state-cycle-signals and the labor lens in state-labor-signals. Raw series: Bureau of Economic Analysis via FRED.
US State Personal-Income Signals (derived)
Quarterly US state personal-income signals from BEA total personal income (FRED <USPS>OTOT series, 49 states + DC — North Carolina excluded, FRED's NCOTOT is corrupted — 1948-Q1 ->): year-on-year growth, quarter-on-quarter change, 4-quarter volatility, 3-sigma anomaly flags, naive-drift forecasts, 5-year growth z-scores, per-quarter growth ranks, each state's share of US personal income, outperform-vs-US and contraction flags. The income lens on US regions — what households actually receive — complementing state-gdp-signals (output), state-labor-signals (unemployment), and state-cycle-signals (coincident index). The US benchmark is the sum of the 50 included state series each quarter (North Carolina excluded — FRED's NCOTOT is corrupted). All rows normalized to country_code USA. Raw series: U.S. Bureau of Economic Analysis via FRED.
Une fois l’allocation gratuite du jour épuisée, les fonctions d’IA peuvent passer par votre propre compte fournisseur.
Conservée uniquement dans cet onglet (effacée à sa fermeture) et envoyée avec chaque requête d’IA. Nos serveurs l’utilisent pour cette requête et ne la stockent ni ne la journalisent jamais.