Catalyst Cooperative

Catalyst Cooperative

@catalyst.coop

We help #climate advocates, policymakers, & researchers working on the #EnergyTransition by liberating #OpenData about the US energy system using #Python based #DataEngineering We're also a worker cooperative.

2.4K followers1.3K following653 posts

Catalyst Cooperative
@catalyst.coop

A good convo with Shreya Shankar, a new professor (starting in 2027) at CMU on the dbt podcast about whether and how and under what circumstances data agents can really do analysis. Spoiler-alert: having clean, well-normalized tables & rich metadata helps a lot! roundup.getdbt.com/p/can-data-a...

roundup.getdbt.comCan data agents do real analysis?Carnegie Mellon Professor Shreya Shankar on why data agents fall short at real analytics work, what a semantic layer changes, and how to evaluate LLM outputs you can trust.
Catalyst Cooperative
@catalyst.coop

With less than 10 members, it makes no sense for Catalyst to run our own open model inference server, but as part of a larger pool of users with similar requirements, it could.

Nathan Schneider@ntnsndr.in

A month ago, I found myself in the unexpected position of founding an AI startup. Today, the Inference Co-op adopted its first real governance charter. We can't make AI good, but we can own our choices, and make them together. Want to join? https://inference.coop

nathanschneider.infoThe Inference Cooperative is democratic!A month ago, I found myself in the unexpected position of founding an AI startup. Today, the Inference Co-op adopted its first real governance charter. We can't make AI good, but we can own our choices, and make them together. Want to join? https://inference.coop

Catalyst Cooperative reposted

Nathan Schneider
@ntnsndr.in

A month ago, I found myself in the unexpected position of founding an AI startup. Today, the Inference Co-op adopted its first real governance charter. We can't make AI good, but we can own our choices, and make them together. Want to join? https://inference.coop

nathanschneider.infoThe Inference Cooperative is democratic!A month ago, I found myself in the unexpected position of founding an AI startup. Today, the Inference Co-op adopted its first real governance charter. We can't make AI good, but we can own our choices, and make them together. Want to join? https://inference.coop
Catalyst Cooperative
@catalyst.coop

Zax looks like a very cool new tool. Bringing a SQL interface to high-dimensional array data in the cloud. Designed to bridge the gap between monster scientific datasets and the often tabular slices and aggregations that get used downstream. www.earthmover.io/blog/compute...

earthmover.ioOur compute roadmap: introducing Zax SQL | EarthmoverWe're introducing Zax, a new query and compute engine for multidimensional scientific data, and Zax-SQL, the first interface to it. Zax plans your computation for you, so you are no longer responsible...
Catalyst Cooperative
@catalyst.coop

New PUDL data release for for September. Updated hourly wind/solar generation profiles, more EQR filings for 2026Q2, some new EIA tables from open source contributors, and much more. We're also deprecating SQLite and moving to @duckdb.org for our DB outputs. #EnergySky github.com/orgs/catalys...

github.comPUDL v2026.9.0 is out; FERC EQR 2026Q2 refreshed · catalyst-cooperative · Discussion #5590This is a regular monthly PUDL data release, primarily motivated by updating the EIA-860M monthly data through August 2026. The biggest change this month is that we’ve started publishing a DuckDB v...
Catalyst Cooperative
@catalyst.coop

The World Sufficiency Lab is hiring: "We work to advance sufficiency as a foundational principle for achieving human well-being within planetary boundaries and for guiding the metamorphosis of economies, institutions, and territories." www.thesufficiencylab.org/news/2026-08...

thesufficiencylab.orgJob Openings | World Sufficiency LabThe World Sufficiency Lab (WSL) is a leading international hub for research, policy engagement, and action on sufficiency. We work to advance …

Catalyst Cooperative reposted

Emily Grubert
@gruberte.bsky.social

Literally the only people who didn’t expect this to be extremely hard to the point of foolishness are the people who decided they were willing to do it, which is exactly the story of risk increasing through transfer to lower info or more risk tolerant owners everywhere else

Sammy Roth@sammyroth.bsky.social

Why Restarting a Nuclear Power Plant Can Be Much Harder Than Expected www.nytimes.com/2026/07/27/b...

nytimes.comWhy Restarting a Nuclear Power Plant Can Be Much Harder Than ExpectedThe first U.S. attempt at reopening a shuttered reactor has been held back by run-down equipment and a lack of planning.
Catalyst Cooperative
@catalyst.coop

We worked through migrating code out of notebooks into shareable, re-usable packages; creating reproducible Python environments, and the basics of data and software testing using data and examples from the energy systems landscape. Looking forward to meeting IRL again! Or at least in office hours.

Catalyst Cooperative
@catalyst.coop

Hey energy nerds, yet another PUDL data release, this time with a bunch of quarterly core dataset updates, including EIA Forms 191, 860, & 923, FERC 714, EPA CEMS, & PHMSA natural gas reports. PUDL v2026.8.0 discussion thread and links here: github.com/orgs/catalys... #EnergySky #OpenData

github.comPUDL v2026.8.0 and FERC EQR 2026Q2 are out · catalyst-cooperative · Discussion #5485This is a regular quarterly PUDL release with updates for many of our core datasets. We've also added a raw version of the EPA MATS dataset (not yet distributed), and have started including Puerto ...

Catalyst Cooperative reposted

Daniel Huppmann
@danielhuppmann.bsky.social

🎉 New release v3.5 of our #opensource python package #pyam-iamc for scenario analysis and data visualization Now supporting central-90%-limits for statistical summaries of ensemble diagnostics - a stepping stone towards the next release & reproducible analysis of the #ScenarioCompass ensemble

github.comGitHub - IAMconsortium/pyam: Analysis & visualization of energy & climate scenariosAnalysis & visualization of energy & climate scenarios - IAMconsortium/pyam

Catalyst Cooperative reposted

Zane Selvans
@zaneselvans.org

On the one hand, I updated our parquet schema checks to be able to run on even the 5-billion row EQR transactions table. On the other hand, this just revealed that none of the primary keys FERC identifies in their data dictionary are actually unique.

Catalyst Cooperative@catalyst.coop

Fresh from the bitmines, new FERC EQR Parquet outputs including 2026Q2 up in our public S3 bucket: s3://pudl.catalyst.coop/ferceqr/ #EnergySky #Electricity #OpenData

Catalyst Cooperative reposted

Jenny Chase
@solarchase.bsky.social

Replying to @sarahtaber.bsky.social

Sorry but this isn't really true - the March solar export spike was due to the Chinese export tax rebate ending as the title mentions, the exports were down again in April and May. We do expect effects, but mostly longer term except in a very few markets.

Catalyst Cooperative
@catalyst.coop

Fresh from the bitmines, new FERC EQR Parquet outputs including 2026Q2 up in our public S3 bucket: s3://pudl.catalyst.coop/ferceqr/ #EnergySky #Electricity #OpenData

Catalyst Cooperative
@catalyst.coop

The PUDL skill is built on top of a lower-level Frictionless Datapackage skill that we made to work with any datapackage metadata -- this is a JSON metadata format originally created by @okfn.bsky.social -- you can see the actual markdown instructions here: www.skills.sh/catalyst-coo...

skills.shdatapackage — catalyst-cooperative/agent-skillsExplore and query any dataset annotated with a Frictionless Data Package descriptor (datapackage.json). Use this skill whenever a user wants to discover what…
Catalyst Cooperative
@catalyst.coop

Then the PUDL skill adds a bunch of stuff that's specific to our metadata -- extensions to the official datapackage standard like physical units, data sources, naming conventions, data quality warnings, how to look up stuff in the PUDL docs, etc: www.skills.sh/catalyst-coo...

skills.shpudl — catalyst-cooperative/agent-skillsExplore and understand PUDL energy data: discover which tables exist, look up column meanings and usage warnings, and load Parquet files from S3 or a local…

Catalyst Cooperative reposted

Matthias Fripp
@matthiasfripp.org

In case you missed it, @brendan.bsky.social, @michelle-solomon.bsky.social and I just published an @energyinnovation.org report showing that the U.S. can save $5 billion per year by meeting load growth with clean power and demand-side flexibility instead of keeping old coal and gas plants running. 🧵

energyinnovation.orgLet The Sun In: Clean Energy Is The Cheapest Way To Meet Rising DemandMeeting U.S. demand growth with fossil fuels will add $29.7 billion to customer bills per year by 2030, clean energy can do it for less.

Catalyst Cooperative reposted

Matthias Fripp
@matthiasfripp.org

Replying to @matthiasfripp.org

This was done using the open-source Switch grid model and a a lot of Python code. All data and code are open-source, available at github.com/EnergyInnova.... Thanks to the PowerGenome, PUDL @catalyst.coop and ReEDS projects for awesome data resources!

github.comGitHub - EnergyInnovation/Switch-US-Load-Growth-2030Contribute to EnergyInnovation/Switch-US-Load-Growth-2030 development by creating an account on GitHub.
Catalyst Cooperative
@catalyst.coop

We call this "raising the floor" -- we're not trying to make the *best* version of a thing, we're trying to make the *free, public version* of a thing better. Which has distributional effects.

Zane Selvans@zaneselvans.org

So they don't need to have the *best* AI, they just need to keep the margin between US rent-seeking AI and Chinese open AI small enough that the whole price-sensitive market (which is maybe everyone...) goes with the open models and the US companies can't make back their CapEx.

Catalyst Cooperative
@catalyst.coop

Another pilot workshop coming up in our series on #Python skills for #energy researchers! Join us online, Tuesday 7/14 2-4p EDT and we'll cover how to use pytest and pdb to make sure your system is behaving. +

Catalyst Cooperative reposted

Jason Scott
@textfiles.com

I'll summarize this: Mamdani asked NYers to keep air conditioning at 78 degrees. A wave criticism hit online. A MSNBC broadcast pointed out the US Department of Energy recommends that 78 degree figure. Immediately after, the US Department of Energy deleted 6,000 web pages related to energy saving.