njudd

outs5SD <- function(vec, n_outsONLY = FALSE){
  m = mean(vec, na.rm = T)
  s = sd(vec, na.rm = T)
  # if less than 5 sd below or above the mean replace with na's
  print(paste0(length(vec[vec < m - s*5 | vec > m + s*5]), " number of outliers above or below 5SD"))
  vec[vec < m - s*5 | vec > m + s*5] <- rep(NA, length(vec[vec < m - s*5 | vec > m + s*5]))
  if(n_outsONLY == FALSE) {
    return(vec)
  } else {} # no return of vector
}

njudd

if (!require(pacman)) {install.packages("pacman")}

pacman::p_load(data.table, psych)

load(url("http://alecri.github.io/downloads/data/dental.RData")) # snagging random long data
setDT(dental)
dental

dental$delta <- dental$y14 - dental$y8
describe(dental$delta)

njudd

cols_to_res <- colnames(cog)[str_detect(colnames(cog), "_T")]

col_res <- function(df, response_var, predictor_string, model_func = lm){
  # predictor_string <- paste(predictor_vars, collapse = " + ")
  # Construct the full formula dynamically
  formula_string <- paste(response_var, "~", predictor_string)
  print(formula_string)
  
  current_formula <- as.formula(formula_string)
  

njudd

#!/bin/bash

# Check if a file name is provided
if [ -z "$1" ]; then
  echo "Usage: $0 <filename>"
  exit 1
fi

# Extract the file name and ensure it has the .R extension
filename="$1"

njudd

library(purrr); library(stringr)

map_chr(str_split(df$col2, ", "), 1) # the number indicates the vec you want based on the split

njudd

pacman::p_load(data.table, stringr)

setDT(ct)

col_fence.s <- c("col1", "col2", "col3")
ct[,(str_c(col_fence.s, ".s")) := lapply(.SD, function(x) as.numeric(as.character(scale(x)))), .SDcols=col_fence.s] # making new ones with a .s suffex

# super sketchy do to only as.numeric incase it is a factor!!! updated to as.numeric(as.character

njudd

pl1 <- ggplot(df, aes(x)) +
  geom_rect(aes(xmin = .8, xmax = 1.2, ymin = 0, ymax = 0.4, fill = "lightblue", alpha = .3)) +
  stat_function(fun = dnorm, n = 200, args = list(mean = 0, sd = 1)) +
  #geom_density(color = "darkblue", size = 1) +
  theme_minimal(base_size = 12) +
  theme(legend.position = "none")  +
  labs(y = "Density of participants (n = 200)", x = "Ability on a Task (SD)") +
  scale_x_continuous(breaks = c(-3, -2, -1, 0, 1, 2, 3), limits = c(-3.5, 3.5))
  
ggsave("~/Desktop/abilitychallange_s1.png", pl1, bg = "white", width = 8, height = 4)

njudd

# Nicholas Judd | Donders Institute
# 26-12-2022
# njudd.com

# replicating https://twitter.com/rogierK/status/1004443199031607296/photo/2
# Wagenmakers, E.-J., & Farrell, S. (2004). 
# AIC model selection using Akaike weights. 
# Psychonomic Bulletin & Review, 11(1), 192–196.

if (!require(pacman)){

njudd

iris |> 
  (\(.) {head(.)})()

# or an example with plotting

diamonds |> (\(.) {
  ggplot(., aes(carat, price)) + geom_point()
  }) () #https://stackoverflow.com/questions/73752006/using-the-base-r-pipe-to-pipe-ggplot2-to-plotly

njudd

# No need to call Bill Gates! 
# here is how you get correlation tables by group with their p-val matries and even plot them
library(Hmisc); library(tidyverse); library(corrplot)

result <- iris |> # base R pipe
  #select() # remove col's you don't want in the correlation matrix
  group_by(Species) |> # grouping col
  group_map(~rcorr(as.matrix(.))) |>
  setNames(unique(sort(iris$Species)))