LightNews LightNews
reciting.bsky.social
@reciting.bsky.social
22 followers 270 following 1 posts
Posts Replies Media Videos
Reposted
shashj.bsky.social
A paper on the probability words used in intelligence analysis. Each word is meant to relate to a specific probability range. But people interpret these words in varied ways. "Participants greatly overestimated the probability for realistic possibility"
www.tandfonline.com/doi/full/10....
March 28, 2025 at 9:50 AM
Reposted
datavisfriendly.bsky.social
ANNOUNCE:
📊 I'm teaching my favorite course, _Psychology of Data Visualization_ again starting in Jan. Among other topics: #history of #dataviz, graph #perception, #rstats graphs, data journalism, ...
Course materials available at: friendly.github.io/6135/
I'll use the tag #psy6135 for posts
Hex logo for my course, Psychology of Data Visualization, featuring an image of various graphic ideas inside a brain
December 3, 2024 at 9:26 PM
Reposted
andrew.heiss.phd
So @jbgruber.bsky.social's {rollama} #rstats package here is actually super super cool
Screenshot from R with this code, running on an Ollama that's running locally through Docker Compose. library(tidyverse) library(rollama) q <- tribble( ~role, ~content, "system", "You assign text into categories. Answer in JSON format with the category, percent confidence, and your reasoning, in this structure: {'category': 'blah', 'confidence': 0.6, 'reasoning': 'blah'}", "user", "text: The food was awful\ncategories: positive, neutral, negative" ) answer <- query(q) #> ✔ llama2 is thinking ⠇ [17.6s] #> #> ── Answer ────────────────────────────────────────────────────────────── #> {'category': 'negative', 'confidence': 0.8, 'reasoning': 'The text explicitly states that the food was awful, which is a negative evaluation.'}
March 2, 2024 at 4:29 PM
Reposted
jperkel.bsky.social
This is so cool. R running in the browser thanks to WebAssembly and {webR}, no installation required! Check it out! #Rstats jperkel.github.io
webr-demo AUTHOR: Jeffrey M. Perkel PUBLISHED: December 14, 2023 WEBR STATUS: 🟢 Ready! WebR This document demonstrates the webR package for the #rstats community. Using it, in conjunction with the quarto-webr Quarto extension, allows users to create Quarto documents with code that runs at runtime instead of during document rendering. p <- mtcars |> ggplot(aes(x = wt, y = mpg, color = factor(cyl))) + geom_point(shape = 21) p We can modify this base graph (stored as the variable p), for instance by changing the background from grey to white. p + theme_minimal() Now, replace the text in the previous cell to read: p + labs(title = "Title", x = "Weight", y = "Miles-per-gallon") + theme_minimal() (You can click the clipboard icon at the upper-right of the cell above to copy its text) and click Run Code again.
December 14, 2023 at 11:43 PM
Reposted
joewdavies.bsky.social
Bivariate maps showing the relationships between Temperature and Precipitation.

Made with Plot, based on a great tutorial by Fil: observablehq.com/@jo...

Data from TerraClimate (1958-present).

#cartography #maps #mapping #climate #dataviz #SciArt
January 6, 2025 at 7:00 AM