North Dakota Industrial Commission (NDIC) Oil & Gas Division data

Oil & Gas Drilling - the state's entire regulatory memory of one basin, kept current. The Oil & Gas Division of North Dakota's Department of Mineral Resources, answering to the Industrial Commission, regulates every well drilled in the Williston Basin - which makes its files the closest thing the Bakken play has to a system of record. What Datadory hands you is that record assembled: 43,824 wells, one row per borehole, each carrying its API number, operator of record, filed name, total depth, spud date, producing field, well type, regulatory status and surface coordinates.

The well inventory is only the spine. Around it sit the companion layers a working Bakken model actually wants: the active drilling rig list naming each rig, its operator, its current location and where it moves next; daily activity reports month by month; a rig statistics workbook tracking counts over time; drilling spacing units, oil fields, gas plants and seismic survey layers; and monthly production reports with historical county and state statistics. One identifier scheme - the state file number matched to the API number - runs through all of it, so a well found in the inventory joins to its unit, its field and its county's output without a reconciliation project in between.

Get a sample of this dataset: pick your counties or operators and rows arrive exactly in the shape documented below.

Three row types carry the record - the well file, today's rigs, and the activity trail:

Well file - one row per borehole
fileno     : 8
api_no     : 33-043-00002-00-00
operator   : PRAIRIE OIL & GAS
well_name  : ARMSTRONG 1
td         : 3884              <- total depth, feet
spud_date  : 1928-05-28
field_name : WILDCAT
status     : DRY
latitude   : 46.96710984
longitude  : -99.86640409
County     : KIDDER

Active rig - one row per rig, current day
rig           : PATTERSON 276
operator      : Phoenix Operating LLC
well_name     : HERCULES 1-12-13 5H
location      : LOT3 1-161N-93W
file_no       : 42559
api           : 33013020070000
start_date    : 2026-08-17
next_location : Undetermined

Active rig - second row, same day
rig           : STONEHAM 18
operator      : PETRO-HUNT, L.L.C.
well_name     : Tachenko 144-98-8D-5-1H
location      : NWNE 17-144N-98W
file_no       : 42328
api           : 33007019600000
start_date    : 2026-08-18
next_location : Undetermined

Read the two halves against each other and the value shows up. The Armstrong 1 row is a century-old wildcat in Kidder County: dry, 3,884 feet, filed before horizontal drilling was a phrase - proof the inventory is a full historical file, not just the shale-era wells people usually mean by "Bakken data." The Patterson 276 row is the present tense, dated August 2026, complete down to where the rig goes after this hole. Same identifier scheme on both, ninety-eight years apart. Spud dates arrive normalized off their raw system timestamps onto real calendar dates, so the parsing pass most regulator extracts demand has already happened.

Fourteen columns are verified against live queries and anchor the dictionary below: the twin identifiers (fileno, api_no), who drilled what (operator, well_name), how deep and when (td, spud_date), where it landed administratively and physically (field_name, status, well_type, county, latitude, longitude, and the legal description - quarter-quarter section, section, township, range and footage calls). The rig list carries its own matching set: rig, operator, well_name, location, file_no, api, start_date, next_location. Every column in the table is live-query verified, not inferred.

Beyond the verified core, the wider service holds layers this page documents by family rather than column-by-column: horizontal wellbore geometries, spacing-unit boundaries, oil-field polygons, gas plants and seismic survey outlines. Those sit under additional fields on request until your sample fixes exact shapes for the acreage you care about.

• Geo: State of North Dakota - which is to say the Williston Basin and the Bakken play as the regulator defines them, not as a vendor's shapefile does. Wells locate to county, to decimal-degree coordinates and to legal description down to quarter-quarter section and footage calls, so mapping and lease work both resolve from the same row.
• Temporal: Spud records run back to 1928 and forward to wells drilling now. The historical file and the current-day rig list are different grains, both included: one answers "what has ever been drilled here," the other answers "who is drilling here this week." Production reports land monthly by county and state.
• Granularity: Individual wells - 43,824 of them - with rigs counted per rig, activity logged per day, and production rolled up at well, county and state level. That mix lets a single extract serve a reserve screen and a statewide activity chart without two sourcing efforts.

The differentiator against multi-state aggregators is depth on one play: no schema compromises to fit twenty regulators, because there is only one, and it has been writing this file since Calvin Coolidge was president.

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

Ask for all 43,824 wells or just your footprint: McKenzie County horizontals only, one operator's full drilling history, or every well spudded into three townships since a date you choose. Extracts arrive shaped to the dictionary above with identifiers intact, so well rows join to their spacing units, their field assignments and their county production without post-hoc reconciliation. Sample first - request the counties, fields or operators you actually cover, check the columns against your own records, then set the cadence. Daily suits an activity monitor; weekly fits a valuation refresh; hourly earns its keep when you are watching a rig move.

• Data scientists get a century-deep, single-schema well panel - the clean base layer for type curves, anomaly detection or spud-date models, minus the scraping-and-repair phase. Patterns live on the data scientists use cases page.
• Investors and quant researchers screen operator footprints and vintage quality straight from the regulatory record - count a company's permitted locations, time its permit-to-spud gaps, rank its counties - without trusting investor decks for either. See investors quants use cases.
• Market researchers and consultants cite figures whose basis is mandated operator filings, and pair the rig list with the production reports to separate drilling activity from actual output. See market researchers use cases.
• Developers and builders wire the well inventory and rig feed into dashboards and land tools; the coordinate-plus-legal-description layout means a map pin and a lease abstract come from the same row. See developers builders use cases.
• Journalists, academics and students trace any figure back to the agency that collected it, and get a ninety-year baseline for asking whether this boom resembles the last ones.
• Land and competitive intelligence teams watch spacing units, field assignments and rig movements county by county - the neighbor's next location is a column, not a rumor.

• Sondio Data Platform - Oil & Gas Wells and Production Data - multi-state breadth over 27 upstream sources; pair it when a Bakken thesis needs context beyond North Dakota's lines.
• BSEE Data Center - Offshore Well, Permit, Production & Platform Data - the federal offshore counterpart, for shops whose remit crosses from basin to Gulf.
• Colorado ECMC (COGCC) Data Portal - the Rocky Mountain regulator's portal, useful when a Niobrara question sits next to a Bakken one.
• GOGI Global Oil & Gas Infrastructure Features Database - infrastructure context - gathering systems, processing - drawn around the wells this file inventories.
• EIA U.S. Crude Oil & Natural Gas Active Well Service Rigs - national rig counts to benchmark North Dakota's share of the drilling pulse against.

Three glossary notes sharpen the vocabulary before you commit: what drilling permits actually certify versus what a spud date records, how well production data differs when cut per well versus per county, and why monthly production statistics lag the months they describe.

Three caveats ride with this record. First, the verified dictionary covers the well file and the rig list; the companion layers (horizontal wellbores, spacing units, fields, gas plants, seismic surveys) are documented by family until a sample fixes their exact columns. Second, the production side lands at county and state rollups in the published reports, so per-well monthly volumes need pairing with another source or a scoped request. Third, the rig list and daily reports are current-day snapshots - they answer "now," not "history," and the rig statistics workbook is the long-run count series instead. None of these blocks a sample; all three get settled before anything depends on them.