Randy Boyes
@randy.pub
Analytics at presagegroup.com | PhD Epidemiology | #julialang #rstats | publications at www.randy.pub | GitHub @rdboyes, currently working on https://github.com/TidierOrg/TidierPlots.jl | 🐦@randyboyes
I'd go with a regex with a lookahead for the character you want, e.g.
October 9, 2025 at 8:45 PM
I'd go with a regex with a lookahead for the character you want, e.g.
It should only be a month or so until he’s good with one wake-up during the night, and then another few weeks until he sleeps through the night. The “every 2 hours” phase is 2-3 weeks long, in my experience
September 14, 2025 at 9:41 AM
It should only be a month or so until he’s good with one wake-up during the night, and then another few weeks until he sleeps through the night. The “every 2 hours” phase is 2-3 weeks long, in my experience
No wonder they were having a blast
August 21, 2025 at 4:17 PM
No wonder they were having a blast
a man with good taste in llms haha :)
August 13, 2025 at 7:00 PM
a man with good taste in llms haha :)
Might be worth trying! I can confirm at least that it CAN work with ellmer, given the right hardware
August 13, 2025 at 6:27 PM
Might be worth trying! I can confirm at least that it CAN work with ellmer, given the right hardware
I’ve been making a concerted effort to delete apps and touch grass recently and it’s good
What happens on the hellscape formerly known as twitter actually doesn’t matter
What happens on the hellscape formerly known as twitter actually doesn’t matter
July 26, 2025 at 5:28 PM
I’ve been making a concerted effort to delete apps and touch grass recently and it’s good
What happens on the hellscape formerly known as twitter actually doesn’t matter
What happens on the hellscape formerly known as twitter actually doesn’t matter
Or maybe a clearer visual:
June 27, 2025 at 10:20 PM
Or maybe a clearer visual:
The challenge in the post: you can detect the seed used if you know or can guess the mechanism (purple spike at seed 123 should have no reason to exist in non-manipulated data):
June 27, 2025 at 10:01 PM
The challenge in the post: you can detect the seed used if you know or can guess the mechanism (purple spike at seed 123 should have no reason to exist in non-manipulated data):
Unless I'm misunderstanding what you want, it does already - you just have to use the ...
June 13, 2025 at 5:15 PM
Unless I'm misunderstanding what you want, it does already - you just have to use the ...
It's not an inherent property of the task, no - e.g. here's an example in julia that does the same thing but has the expected variance in the last digits
June 9, 2025 at 12:47 PM
It's not an inherent property of the task, no - e.g. here's an example in julia that does the same thing but has the expected variance in the last digits
Trying out some old Vizz Buzz challenges in TidierPlots - is there any way in #makie to easily add small images? The original version had flags as labels...
![Code to recreate a data visualization showing the medal progress at the 2018 olympics, reproduced below:
using TidierPlots, TidierFiles, TidierData, DataFrames, Dates
wo = read_csv("data/bump (1).csv")
wo_plot = @chain wo begin
@mutate total = gold + silver + bronze
transform(:date => ByRow(x -> 1 + (x-Date("2018-02-10")).value) => :day)
@group_by date
@mutate rank = 1:n()
@ungroup
end
top5 = @chain wo_plot begin
@filter country in ["NED", "GER", "CAN", "USA", "NOR"]
end
final = @chain top5 begin
@filter day == 16
end
ggplot(wo_plot) +
geom_point(@aes(x = day, y = rank, group = country), alpha = .4, size = 15, color = :lightgrey) +
geom_line(@aes(x = day, y = rank, group = country), alpha = .4, linewidth = 6, color = :lightgrey) +
geom_point(data = top5, @aes(x = day, y = rank, color = country), size = 15, alpha = 0.5) +
geom_line(data = top5, @aes(x = day, y = rank, color = country), linewidth = 6, alpha = 0.5) +
geom_point(data = top5, @aes(x = day, y = rank), color = :white, size = 5) +
scale_y_reverse() +
lims(y = c(0.5, 10.5), x = c(0, 18)) +
labs(title = "Pyongchang 2018 Olympic Winter Games",
subtitle = "Countries ranked by overall medals after each competition day",
x = "Competition Days with Medals", y = "Rank") +
scale_color_manual(values = c(:red, :black, :orange, :green, :blue)) +
guides(color = :none) +
theme_minimal() +
theme(
xgridstyle = :solid,
xgridcolor = :lightgray,
xgridvisible = true,
xticks = 1:16,
yticks = 1:10,
leftspinevisible = false,
bottomspinevisible = false,
titlealign = :left,
titlesize = 24,
subtitlesize = 16
)](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:2h5e6whhbk5vnnerqqoi256k/bafkreiducjpjc7n2rtbbuho42y6qnlcvqpex6kwnosctssvssqemvsb4zy@jpeg)
June 4, 2025 at 11:08 PM
Trying out some old Vizz Buzz challenges in TidierPlots - is there any way in #makie to easily add small images? The original version had flags as labels...
They're "hack-proof and secure", obviously
May 2, 2025 at 9:02 PM
They're "hack-proof and secure", obviously
I'm bookmarking this for reference since R's NSE stuff always throws me off!
One of the nice things about working in @tidierjl.bsky.social is that you can drop down into #julialang at any time, which has a nice division between "normal evaluation" and "shenanigans":
One of the nice things about working in @tidierjl.bsky.social is that you can drop down into #julialang at any time, which has a nice division between "normal evaluation" and "shenanigans":
![Julia source code example:
# One-liner if you're willing to write column names as symbols
count_ones(data, c...) = combine(data, [c...] .=> x -> sum(x .== 1) .=> [c...])
count_ones(mtcars, :VS, :Carb)
# 1×2 DataFrame
# Row │ VS Carb
# │ Int64 Int64
#─────┼──────────────
# 1 │ 14 7
# macro version to allow unquoted names
macro count_ones(data, c...)
return Expr(:call, :count_ones, data, (Meta.quot.(c))...)
end
@count_ones(mtcars, VS, Carb)
# 1×2 DataFrame
# Row │ VS Carb
# │ Int64 Int64
#─────┼──────────────
# 1 │ 14 7](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:2h5e6whhbk5vnnerqqoi256k/bafkreicf2ppjey5vw3e2jinabihzovg3tlp6fx7kkxpyweud6ikdifamuy@jpeg)
April 25, 2025 at 12:45 PM
I'm bookmarking this for reference since R's NSE stuff always throws me off!
One of the nice things about working in @tidierjl.bsky.social is that you can drop down into #julialang at any time, which has a nice division between "normal evaluation" and "shenanigans":
One of the nice things about working in @tidierjl.bsky.social is that you can drop down into #julialang at any time, which has a nice division between "normal evaluation" and "shenanigans":
You can cut one more pixel if you're willing to assume you can interpolate to get blank points
April 23, 2025 at 7:25 PM
You can cut one more pixel if you're willing to assume you can interpolate to get blank points
Cut the intermediate braille labels for even more information density! (people should of course be able to infer that the categories are in alphabetical order between a and g ;) )
April 23, 2025 at 4:23 PM
Cut the intermediate braille labels for even more information density! (people should of course be able to infer that the categories are in alphabetical order between a and g ;) )
The one piece of good news in the portfolio today
April 10, 2025 at 8:56 PM
The one piece of good news in the portfolio today
"I could only process 60k rows before my hard drive overheated"
March 14, 2025 at 1:37 PM
"I could only process 60k rows before my hard drive overheated"
TidierPlots 0.11 is available without removing or uninstalling anything - on first pass it installs 0.10 but running Pkg.update() upgrades it to 0.11!
March 7, 2025 at 8:30 PM
TidierPlots 0.11 is available without removing or uninstalling anything - on first pass it installs 0.10 but running Pkg.update() upgrades it to 0.11!