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.
2 jeux de données
US parcel carriers 2026-27 peak surcharges (agent-curated)
Derived daily panel (2026-09-26..2027-01-17, per carrier): on each date, how many surcharge lines are active, the headline standard-residential per-package fee in effect (NULL for USPS, which raises rate tables instead — flagged by rate_increase_active), whether the date is in the carrier's peak-of-peak window, and a documented 0-100 peak_cost_pressure = round(100 * (0.6 * active_line_count / carrier_line_total + 0.4 * residential_component)) where the residential component is fee / carrier_max_fee (UPS 0.75, FedEx 0.80, Amazon 0.75, OnTrac 1.00) or rate_increase_active for USPS. A shop joins this on ship date + carrier to budget the seasonal per-package premium; 570 rows, deterministic in code from the curated window table. Annual snapshot, yearly cadence.
Agent-curated cost intelligence: 75 surcharge-window rows covering UPS, FedEx, Amazon Shipping, OnTrac, and USPS for the 2026-27 peak season (2026-09-26..2027-01-17). Each row gives the carrier, charge line (per-package demand, accessorial demand, or USPS rate increase), the exact pricing window (shoulder / peak / close), the USD amount, and a per-row verification level (UPS read verbatim from the carrier's official schedule; FedEx and USPS cross-verified across independent sources; Amazon Shipping from four agreeing analyses; OnTrac from a single compiled table). Join to shop orders by ship date + carrier for Q4 margin and shipping-cost forecasting. Provenance: agent-curated (intel-3 run 2026-09-30) from the carriers' published schedules; the code encodes the schedules verbatim and derives ids and hashes. Caveats: high-volume (>20k/week) tiered tables excluded (enterprise only); USPS stacks on the April 2026 temporary 8% change; retail USPS amounts excluded (commercial table encoded).
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.