The US Energy Information Administration (EIA) releases a weekly petroleum status report on Wednesdays at 10:30 AM EST, or on Thursdays at 11:00 AM EST in case of weekly US holidays (www.eia.gov).
The most closely followed portion of the report is the one covering US commercial stocks of crude oil and petroleum products.
Crude oil and petroleum products demand, production, import and export are also reported on a weekly basis.
The figures in the report are ingested by dSystems.ai data pipeline as soon as released and added to the respective time series database entries.
Forecasts are made for all the series, so that there is an expected release number, and a related surprise number after the actual release of each series weekly datapoint.
Given the strong seasonality displayed by most of these time series, Econometric models such as univariate Auto-Regressive Integrated Moving Average (ARIMA), and multivariate Vector-Auto-Regressive (VAR) and Vector-Error-Correction (VEC) are typically used as forecasting tools for these series, alongside standard linear regression approaches.
The forecasts reported on dSystems.ai are generated using proprietary Deep Learning models.