[OC] Sources of Utility-Scale Power Generation in the US

Here are some maps about the sources of utility-scale power generation and a slide on net electricity balance to add context. There are two important caveats:

1. There is a small amount of distributed generation that isn’t captured in the data. This is probably responsible for Hawaii’s small deficit in electricity production in this data. It isn’t able to import electricity but that difference is likely made up by rooftop solar. In most states, this amount is negligible.
2. There is a significant amount of trading of electricity between states. I included the final slide on net electricity balance because some states have electricity supplies that look cleaner than they are because they import dirtier energy from other states. For example, when I was growing up in Seattle there was some controversy around Puget Sound Energy buying electricity from a coal power plant in Montana. I suspect something similar is happening in my current home state of Virginia, which needs to import (on net) around 28% of its electricity. While Virginia generates very little power with coal, West Virginia is a net exporter of electricity which relies heavily on coal.
3. Another less important note on net electricity balance is that it doesn’t account for losses in transmission, which is a major reason for why the state-level surpluses add up to more than the deficits.

Data: US Energy Information Administration API

Tools: R (packages: dplyr, ggplot2, sf, usmap, tools, ggfx, grid, scales, cowplot, showtext, sysfonts, colorspace)

library(httr)
library(jsonlite)
library(tidyverse)

get_eia <- function(endpoint, api_key, params = list(), years = NULL) {
base <- paste0(“https://api.eia.gov/v2/”, endpoint)
if (is.null(years)) {
res <- GET(base, query = c(list(api_key = api_key), params))
json <- fromJSON(content(res, “text”), flatten = TRUE)
return(json$response$data)
}
results <- lapply(years, function(y) {
message(“Pulling year: “, y)
query <- c(
list(api_key = api_key),
params,
list(start = y, end = y)
)
res <- GET(base, query = query)
json <- fromJSON(content(res, “text”), flatten = TRUE)
df <- json$response$data
if (!”period” %in% names(df)) {
df$period <- y
}
Sys.sleep(0.2)
return(df)
})
out <- bind_rows(results)
return(out)
}

gen_all <- get_eia(
“electricity/electric-power-operational-data/data/”,
api_key,
list(
frequency = “annual”,
`data[0]` = “generation”,
`facets[sectorid][]` = “98” #Electric Power = 98
),
years = 2024
)

cons <- get_eia(
“electricity/retail-sales/data/”,
api_key,
list(
frequency = “annual”,
`data[0]` = “sales”
),
years = 2024
) #retail sales

Could we get the last map as a portion of the total power usage? Feels like that might be more useful in some ways

I thought Texas had an isolated power grid? How is it an exporter?

Iowa is impressive. At least all that wind is useful.

Seems like California could use some more reactors.

I would be really interested to see PA’s mix 20 years ago…crazy that they are down to just 5% coal. I know they had several large natural gas plants come online the last 10 years or so.

Wtf Florida. The place should be nearly fully solar.

And the irony of them being so vulnerable to climate change.

I’d love to see the same maps, but for each grid

Crazy to see LA so high on gas and so low on hydro when they have so much water flowing through the state constantly

How does this count [Intermountain Power Plant](https://en.wikipedia.org/wiki/Intermountain_Power_Plant) which is located in Utah, but operated by Los Angeles DWP.

How there isn’t more solar rolled out nation wide is baffling. Easily one of the biggest energy policy bungles in our nations history that will continue to set us back decades.