EIA One
@EIA_One
Followers
5K
Following
72
Media
34
Statuses
383
The official account of the Administrator of the Energy Information Administration.
Washington, DC
Joined June 2022
To address emerging energy issues, @EIAgov needs a next generation model that is modular, flexible, transparent, and robust. We are committed to a transparent model development process. If you’re a modeler, please consider providing feedback on this early stage prototype.
We've published our BlueSky prototype—the first step in creating our next-generation model. Check out our website to find a link to the prototype and to learn how to share feedback about the overarching model framework and code implementation. ▶️
3
1
12
If you have a chance to experiment with the tool, let us know how we can improve it and expand the data coverage.
2
0
5
Caveats:.1. This is a python-based Dash app that runs on your local machine. We don't have a live dashboard on the website. Maybe someday. 2. We're pulling past projections from our API, which only goes back to 2015. 3. Only key market indicators are currently included.
1
0
7
It's important for modelers to examine the accuracy of their past projections. Accurate past performance is no guarantee for the future, as structural elements of the system can change rapidly. But we can still learn a lot.
1
0
4
In an effort to increase transparency, we have developed a python-based benchmarking tool to compare our AEO projections against real world outcomes. It includes all past cases and allows you to visualize the pathways. It's available in GitHub:
1
1
10
When we release the AEO, there's always plenty of discussion about the quality of our projections. Those discussions should be informed by a rigorous look at our our past performance.
1
0
1
Producing high quality long-term energy projections is hard. Uncertainty looms large, and there are inevitable surprises that our models do not capture.
1
0
5
For all the long-term modelers out there: we just released an open source benchmarking tool that allows you to plot projections from past editions of the Annual Energy Outlook (AEO) against actual outcomes. Let me explain why this is important. 🧵.
Meet dash-benchmark, a new open-source analytic tool on our GitHub page. Dash-benchmark allows you to visualize projection results and historical values from our Annual Energy Outlook. Check out dash-benchmark and our other open-source tools on our Github page. ⬇️.
5
20
106
Update: battery electric vehicle sales represented 7.1% of the U.S. light duty vehicle market in 2Q24, about the same as last year. Today in Energy (8/26/24) 👇.
eia.gov
Battery electric vehicles (BEVs) made up 7.0% of U.S. sales last year, compared to 5.6% in 2022. Under that data are a few trends that help explain the state of the market and where it might be headed. 🧵👇.
0
1
3
That doesn’t mean our data are perfect. Our folks continuously review the reported data to address concerns about its accuracy and comprehensiveness.
1
0
4
We're not more right or wrong here; it’s just a difference in accounting.
1
0
1
The key point is that we’re not somehow missing large chunks of capacity. It’s either in an operational or planned phase. Our definition simply means that more capacity might show up on the right panel rather than the left relative to ERCOT or other organizations.
1
2
6
We consider generators that are synchronized but not in commercial operation to be in a testing status and count them as “planned” capacity. You can find those units in the 860M under the “Planned” tab. Look for units with "TS" as their status in column U.
1
0
2
On the 860M, we count capacity that is *in commercial operation* to be operational. Some organizations consider generators that have been synchronized but not yet brought into commercial operation to also be operational.
1
0
2
I wouldn’t call differences in EIA / ERCOT data a “discrepancy.” We categorize capacity differently and try to make clear what we’re measuring. In the case of the EIA-860M data, you can find definitions of what we’re counting in the survey instructions:
1
0
2
Thanks for bringing this issue back up, @tylerhnorris. Let’s dig in.🧵.
@douglewinenergy @SimonMahan @mattbuford @FERC @SPGMarketIntel EIA says it doesn't include "synced-not-approved" generators, but ERCOT does in part because synchronized plants can often dispatch at full output, even if they haven't achieve full reg approval. I'm told ERCOT has had operating facilities in "synced-not-approved" category for.
1
0
13