This is a monthly PUDL data release, primarily motivated by updating the EIA-860M monthly data through February 2026. As usual, it also includes all of the other changes that have accumulated on main since our last release.
This month, we have the belated EPA CEMS update for 2026Q1, the annual update for FERC 1, some great community contributions for RUS7 and EIA-176, and an assortment of datapackage, Dagster, and deployment notification improvements.
\n\n\nOverhauled PUDL’s Frictionless Data Package output to conform to the v2 spec. The pudl_datapackage Dagster asset now generates datapackage.json directly during the ETL, including full column types, constraints, and foreign key relationships for every Parquet table. The descriptor is distributed as pudl_parquet_datapackage.json at the top level of the S3 bucket and on Zenodo, allowing potential users to browse the PUDL schema without downloading any data. The pudl_parquet.zip archive also contains a datapackage.json descriptor so it can be used as a self-describing Frictionless package after extraction. A reusable valid_datapackage_check() factory is now available in pudl.dagster.asset_checks to add frictionless v2 validation as an asset check on any datapackage output. See issues #5122, #5237 and PR #5270, #5343. Also makes progress towards catalyst-cooperative/agent-skills#14
Added a bare-bones datapackage for DBF SQLite outputs. See issue #5200 and PR #5275.
\nAdded core_eia176__yearly_gas_supply, which contains cleaned company-level natural and supplemental gas supply data from Part 4 of the EIA-176 survey. See #4711 and #5227.
\nAdded core_eia176__yearly_liquefied_natural_gas_inventory, a new table containing annual LNG storage volume and capacity reported by operators on EIA Form 176 Part 5. Data covers 2002-2024 and includes LNG terminal and marine terminal records. See issue #4695 and PR #5219.
\nAdded 2025 data from FERC form 1. This update includes several new renewable and energy storage fields in several tables. See issue #5214 and PRs #5236, #5325.
\nUpdated the bulk EIA Electricity API data used to fill in redacted fuel prices. See PR #5292.
\nUpdated the raw FERC Form 2 and 6 archives to include 2025 data. This data is converted to SQLite, but not deeply integrated into PUDL. See PR #5292.
\nAdded a data source page for EIA-191. See PR #5267 and issue #4756.
\nUpdated the EIA-930 column descriptions to note that starting in 2024Q3 EIA began reporting more granular renewable energy source categories, differentiating wind and solar plants with and without energy storage, splitting pumped hydro from conventional hydro, and adding new battery storage and geothermal categories. See issue #5335 and PR #5336.
\nAdded validations to RUS7 service interruption tables to ensure subcomponents sum to the total for annual observation periods. See issue #5285 and PR #5286.
\nRenamed the fuel_consumed_mmbtu column in the out_eia923__fuel_receipts_costs, out_eia923__monthly_fuel_receipts_costs, and out_eia923__yearly_fuel_receipts_costs tables. This column is the result of dividing total_fuel_cost by fuel_received_mmbtu. The name fuel_consumed_mmbtu was misleading because the fuel received in these tables is not necessarily consumed in the same month, and the fuel cost is not necessarily associated with fuel received in the same month. The new name, fuel_received_mmbtu, more accurately reflects what the column actually contains. See PR #5294.
Fixed a bug in the Zenodo Data Release script which was not actually skipping top-level directories when deciding what to upload to Zenodo, which caused release failures once we started leaving the ferc*_xbrl directories on the filesystem. See PR #5254.
Refactored Dagster-managed path handling to use a dedicated pudl_paths resource instead of constructing pudl.workspace.setup.PudlPaths directly throughout assets, IO managers, and tests. This makes path resolution more explicit in Dagster contexts and allows interactive definitions to override pudl_input and pudl_output directly when calling pudl.dagster.build.build_interactive_defs(). See PR #5261, #5288.
Added a PUDL devcontainer configuration to make it easier for contributors to get up and running, and to enable the safe use of coding agents in YOLO mode. See PRs #5260, #5287.
\nCleaned up PUDL’s default Dagster wiring by separating default resources from IO managers, giving shared data-config resources clearer defaults, and simplifying the FERC SQLite IO manager and provenance stack. Consolidated the FERC EQR deployment helper assets with the rest of the Dagster package layout. Created a new Dagster definition builder for use in notebooks and other interactive environments outside of a dg-spawned environment: pudl.dagster.build.build_interactive_defs(). See issue #5118 and PR #5242.
Migrated build and deployment notifications from Slack to Zulip. All GitHub Actions workflows that previously posted to Slack now send notifications to the Catalyst Cooperative Zulip instance via the zulip/github-actions-zulip action. A new ZulipNotificationResource Dagster resource was added to send Zulip stream messages from within assets, with best-effort error handling. The FERC EQR deployment helpers in pudl.dagster.assets.deploy.ferceqr were updated to use it. Notification coverage was also expanded to include community activity (issues, discussions, comments, and pull requests from non-Catalyst contributors). See PRs #5298, #5328, #5331.
FERC provenance metadata (Zenodo DOIs, data years, XBRL extractor version) is now stored in the FERC SQLite datapackage files rather than only in Dagster asset metadata. The ferc_to_sqlite asset can now optionally download and reuse pre-built FERC SQLite outputs from the most recent nightly build, skipping expensive re-extraction when the inputs haven’t changed. Set PUDL_FERC_FORCE_EXTRACT=true to force re-extraction regardless. See issue #5220 and PR #5264.
Migrated hashtag-prefixed comments from soon-to-be-machine-generated dbt schema files into their corresponding human-editable schema input files (dbt/schema_inputs/**/schema.human.yml) to preserve their content, since any regenerated schemas will forcibly strip out hashtag comments. See PR #5310.
If you're using PUDL, we would love to hear from you! Even if it's just a note to let us know that you exist, and how you're using the software or data. Here's a bunch of different ways to get in touch:
\nThis community is meant to house the official releases of resources produced by Catalyst Cooperative.
", "description": "Resources generated and archived by the Catalyst Cooperative, mostly related to open energy data, especially pertaining to the US electricity system.", "title": "Catalyst Cooperative", "type": { "id": "organization" }, "website": "https://catalyst.coop" }, "revision_id": 1, "slug": "catalyst-cooperative", "updated": "2023-10-13T15:20:15.304590+00:00" } ], "ids": [ "14454015-63f1-4f05-80fd-1a9b07593c9e" ] }, "id": "3653158", "pids": { "doi": { "client": "datacite", "identifier": "10.5281/zenodo.3653158", "provider": "datacite" } } }, "pids": { "doi": { "client": "datacite", "identifier": "10.5281/zenodo.20766264", "provider": "datacite" }, "oai": { "identifier": "oai:zenodo.org:20766264", "provider": "oai" } }, "revision_id": 4, "stats": { "all_versions": { "data_volume": 31755022999956.0, "downloads": 31950, "unique_downloads": 30178, "unique_views": 16668, "views": 18512 }, "this_version": { "data_volume": 157598625004.0, "downloads": 299, "unique_downloads": 298, "unique_views": 57, "views": 61 } }, "status": "published", "swh": {}, "updated": "2026-06-19T19:44:43.818136+00:00", "versions": { "index": 27, "is_latest": true } }