BLUE

tj mahr

@tjmahr.com

data scientist studying how kids learn to speak, dad, jump roper, bayesian, tjmahr.com

1k followers371 following1.8k posts

Overview Posts Replies

TMtjmahr.comOct 1, 2024 3:43am

that sucked

Chart of jumps per day showing me binging to get 7000 in the last two days to squeak past my monthly jump goal

TMtjmahr.comSep 30, 2024 11:40pm

- the most popular Magic The Gathering format was a community created one, managed by a Rules Committee outside of Wizards of the Coast - last week, RC announced major bans that tanked value of key cards in the format and spurred a massive backlash - this week, RC is dissolved and taken over by WOTC

TMtjmahr.comSep 30, 2024 6:32pm

not a good idea to ask chatgpt for help naming variables chatgpt.com/share/66faee...

ChatGPT

A conversational AI system that listens, learns, and challenges

TMtjmahr.comSep 30, 2024 5:42pm

the new tar_assign({}) will convert `x <- tar_target(command)` into `tar_target(x, command)` so you can even do something like `downstream <- upstream |> f() |> g() |> tar_target()`

$r code targets_main <- tar_assign({ database <- tar_file("data/data.duckdb") data_vss_4_6_8_10 <- tar_file_read( "data/vss-4-6-8-10-ratings.csv", readr::read_csv(!!.x) ) data_vss_12_14 <- tar_file_read( "data/vss-12-14.csv", readr::read_csv(!!.x) |> janitor::clean_names() |> rename( child = subject_initials, vss = viking_researcher_rating, rater = viking_researcher_name, vss_inter = viking_researcher_rating_2, rater_inter = viking_researcher_name_2 ) |> mutate( age_group = ifelse( visit_group == "VSS Age 12", 12, 14 ), vss = vss |> utils::as.roman() |> as.numeric(), vss_inter = vss_inter |> utils::as.roman() |> as.numeric(), ) ) data_vss <- data_vss_4_6_8_10 |> bind_rows(data_vss_12_14) |> select( subject_num, visit_id, child, cohort, visit_num, age_months, age_group, vss, rater ) |> tar_target() })$

TMtjmahr.comSep 30, 2024 5:24pm

i got some ratings as roman numerals today so fyi here's how you convert them to numbers #rstats#> [1] 1 2 3 4 3 2

Reposted by tj mahr

CTcarlbergstrom.comSep 27, 2024 7:46pm

“Sure, the paper on the benefits of preregistration wasn’t preregistered and sure it was outcome-switched and sure it took an ECR a year to sort it out and sure he got doxxed for his troubles, but this just proves its point bc w/o preregistration we wouldn’t know it was flawed” is a hell of a take.

TMtjmahr.comSep 30, 2024 4:40pm

this is uhhh nine hours long youtu.be/DCRfK6Fs0DU?...