BSEE Data Center - Offshore Well, Permit, Production & Platform Data

Oil & Gas Drilling - the federal system of record for everything drilled, produced and built in American federal waters. The Bureau of Safety and Environmental Enforcement's data center is where offshore energy activity on the US Outer Continental Shelf gets reported under regulation, then published: wells and drilling permits, monthly production, platforms, pipelines, leases and the companies behind them. This entry covers the whole center rather than one artifact.

The Well section alone is a permitting archive most vendors would sell as three products: an Application for Permit to Drill (APD) query, an API Lookup with API Number Changes tracking identifier reassignments, a Borehole Query resolving individual boreholes beneath each API number, Bottomhole Pressure Surveys, directional and velocity surveys, well completions, well test PDFs going back to 1975, and eWell submissions spanning APD, APM, End of Operations Reports and Well Activity Reports.

Production runs deeper than most commercial feeds bother to license: OGOR-A at well level, OGOR-B at lease level and OGOR-C for product sales and inventory, monthly from 1996 to current across Gulf of America, Pacific and Alaska planning areas - plus production-by-platform and planning-area cuts, operator rankings by oil and gas, deepwater summaries and Facility/Measurement Point listings.

Then the infrastructure ledger. Platform Masters, structures, locations and removal records; pipeline masters, permits, ROW descriptions, histories and locations; leasing tables covering assignments, lease owners, area/block lookups, serial register pages, decommissioning cost estimates and OSFR; company detail with incidents of non-compliance; reserves estimates through 2023; G&G studies and paleontological reports; and about 55 GIS layer archives mapping blocks, platforms, pipelines, protraction polygons and the continental shelf boundary itself. It anchors our oil gas drilling data hub as the only slice primary that covers federal offshore waters end to end.

Get a sample of this dataset: name the leases, blocks or planning areas and rows arrive in exactly the shape documented below.

Two rows, straight from the source files, tell the story of both halves of the catalog:

OGOR-A production line (ogora2024delimit.txt)
Lease Number   : " 00016"
Area Code      : " E002"        Block       : "00"
Prod Month     : "202401"       Prod Interval: "O" / "Q04"
Oil            : "177190022500"  Days On     : "12"
Operator       : COX OPERATING LLC
First Prod Date: 19720331

Borehole record (5010.txt)
API Number     : 012032016500   Borehole ID : ST00BP00
Well Name      : WILD           Spud Date   : 19840924
Surface Hole Lo: 03580N07118W (MO 0827)
Bottomhole Loc : 03707N07080W (MO 0827)
Longitude      : -87.958630     Latitude    : 30.188450
Spud-to-Compltn: 19850310 -> 19930309

Read the anatomy. The production row is one lease-month cell: who operates it (COX OPERATING LLC, named outright - no anonymized operator codes), where it sits on the OCS grid (area E002, quoted with padding because these are fixed-position fields), which month it produced (202401), and volumes beside the days that produced them, which is how uptime-adjusted rates get computed rather than guessed. The borehole row is one physical hole in the seabed: pinned twice over by survey text and by decimal coordinates, tied to its parent API number, carrying spud and completion dates forty-eight years after the lease's first production date of 19720331. Live rows for whatever leases and months you name ship with your sample.

Three keys recur across nearly every product family and are verified against inspected rows; the rest of the dictionary follows them. The design choice worth noticing: every record is geolocated twice - once by the OCS administrative grid (lease number, protraction area code, block number) and once by physical measurement (survey text positions or decimal coordinates). That redundancy is why offshore analysis joins cleanly against maps without a geocoding project in the middle.

• Geo: The entire US Outer Continental Shelf - Gulf of America (Gulf of Mexico) planning areas, plus Alaska, Pacific and Atlantic waters - always at lease/block/well detail, never regional totals. Mapping layers extend the geometry to protraction polygons and the continental shelf boundary.
• Temporal: Well tests from 1975 to present; OGOR production monthly from 1996 to current; reserves and sands atlases through 2023; velocity surveys limited to those 10 years old or older. A PPDM data-model transition on July 25-26, 2020 means some values can differ from pre-2020 file versions.
• Granularity: Well/borehole, lease, block, platform, pipeline and company level, with monthly production periods - and platform-level production cuts that let one operator's output be separated from their hosts'.

That footprint has no onshore analogue in the slice: state regulators match the depth but stop at their borders, while this reaches across all federal waters from one schema.

API, files, or your warehouse. Daily, weekly, or hourly.

Ask for the whole OCS ledger or just your acreage - one operator's wells since first production, a single planning area's monthly OGOR series since 1996, or platform masters joined to removal records for a decommissioning screen. Extracts arrive shaped to the dictionary above with lease numbers and API numbers intact, so cross-table joins work on arrival instead of after a reconciliation project. Sample first: request the leases, blocks or months you care about and check them against your own operations records before anything else happens.

• Data scientists treat it as the training corpus for offshore activity models - five decades of permits, bores and production under one schema. Patterns live on the data scientists use cases page.
• Market researchers and consultants anchor client deliverables on regulator-collected figures whose basis is statute rather than sampling - the difference between citing and estimating. See market researchers use cases.
• Journalists and academics cite it because every number traces to a mandated operator filing, from a single deepwater spud to fleet-wide incident-of-non-compliance counts.
• Investors and quant researchers read permit filings and monthly production as coincident signals for offshore drillers and service exposure. The investors quants use cases page shows the workflow.
• Developers and builders wire flat tabular extracts into rig-tracking dashboards and supply-chain tools. See developers builders use cases.
• Competitive intelligence and sales teams track which operators hold which blocks, run which platforms, and report what - footprint you can quote row-for-row.

• BSEE Well API (multiregional online query) - the sibling record that dissects one artifact in depth: the 57,482-record well-header file. Start here for breadth, there for depth on wells.
• Texas Railroad Commission Oil & Gas Data Sets - the onshore depth play: weekly Full Wellbore files at county detail. Federal waters versus Texas land, side by side.
• Colorado ECMC (COGCC) Data Portal - 124,387 wells in daily-refreshing shapefiles; the Rocky Mountain counterpart when coverage crosses onto state land.
• EMODnet Human Activities Oil & Gas Wells - European offshore well points; the transatlantic mirror for basin comparisons.
• GOGI Global Oil & Gas Infrastructure Features Database - global infrastructure footprints that put the Gulf's platform density in worldwide context.

Three glossary notes sharpen the vocabulary before you commit: what drilling permits actually certify, how monthly production statistics are compiled from operator reports, and why well production data differs when reported at well versus lease level. The regulatory-versus-ML tradeoff continues on the comparison page.

Three caveats ride with this record. Field definitions carry an inferred-confidence tag because per-file dictionaries travel with the raw files rather than as web tables - parse fixed-width variants only after checking their layout guides. No published per-product refresh schedule exists, so cadence is observed rather than promised. And the July 2020 PPDM model switch means pre- and post-2020 vintages can differ in some columns. None of these blocks a sample; all three get settled against delivered rows before anything depends on them.