trinker

# https://mathworld.wolfram.com/LookandSaySequence.html

library(ggplot2)
library(stringi)

tic <- Sys.time()
len <- 68
s <- rep(NA, len)
s[1] <- 1
tm <- rep(NA, len)

trinker

#' Transform Model Output for Use with the LDAvis Package
#'
#' Convert a \pkg{topicmodels} output into the JSON form required by the \pkg{LDAvis} package.
#'
#' @param model A \code{\link[]{topicmodel}} object.
#' @param \ldots Currently ignored.
#' @seealso \code{\link[LDAvis]{createJSON}}
#' @export
#' @examples
#' \dontrun{

trinker

prob <- unlist(lapply(1:1000000, function(i){
    candles <- sort(runif(2, 0, 1))
    cut  <- runif(1, 0, 1) 

    (cut > candles[1]) & (cut < candles[2])
}))

mean(prob)

trinker

In this post I have a few goals:

1. Become (re-)familiar with available geoms
2. Become (re-)familiar with aesthetic mappings in geoms (stroke who knew?)
3. Answer these questions:
<ul>
	<li>How often do various geoms appear and how often do they have required aesthetics?</li>
	<li>How often do various aesthetics appear and how often are they required?</li>
	<li>What geoms are most similar based on mappings?</li>
</ul>

trinker

## https://youtu.be/094y1Z2wpJg

library(tidyverse)

collatz <- function(x){

    v = c(x)
    i = 1
   
    while (v[i] != 1){

trinker

#' Find Optimal Number of Topics
#' 
#' Iteratively produces models and then compares of the harmonic mean of the log likelihoods in a graphical output.
#' 
#' @param x A \code{\link[tm]{DocumentTermMatrix}}.
#' @param max.k Maximum number of topics to fit (start small [i.e., default of 30] and add as necessary).
#' @param burnin Object of class \code{"integer"}; number of omitted Gibbs iterations at beginning, by default equals 0.
#' @param iter Object of class \code{"integer"}; number of Gibbs iterations, by default equals 2000.
#' @param keep Object of class \code{"integer"}; if a positive integer, the log-likelihood is saved every keep iterations.
#' @param method The method to be used for fitting; currently \code{method = "VEM"} or \code{method= "Gibbs"} are supported.

trinker

general_rescale <- function(x, lower, upper){

    rng <-  range(x, na.rm = TRUE, finite = TRUE)
    if (diff(rng) == 0) return(stats::setNames(rep(upper, length(x)), names(x)))
    (x - rng[1])/diff(rng) * diff(range(c(lower, upper))) + lower

}

x <- c(NA, 1:10)

trinker

text = ['df dft sdf', 'sd fdggg sd dfhhh d', 'ddd']

def dbllttrwordrev(match):
    match = match.group()
    return '<<{}>>'.format(match[::-1])   

{
    'function': [re.sub("\\b\\w*([a-z])(\\1{2,})\\w*\\b", dbllttrwordrev, x, flags = re.IGNORECASE) for x in text],

    'lambda': [re.sub("\\b\\w*([a-z])(\\1{2,})\\w*\\b", lambda x: '<<{}>>'.format(x.group()[::-1]) , x, flags = re.IGNORECASE) for x in text]

trinker

###############################
## Plotting Likert Type Data ##
###############################
##------------------------------------------------------------------------
## Note: Plotting horizontal stacked bar plots in ggplot2 with Likert type
## data is a non-trivial task.  Stacking is not well defined for mixed
## negative/positive values on a bar.  This requires splitting the data
## set into two different parts (positive/negative), plotting each side
## separately, and filling the colors manually.  This script adds complexity
## for neutral scales.

trinker

## Load dependencies
library(quanteda)
library(sentimentr)
library(tidyverse)
library(lexicon)
## Data set from sentimentr package
dat <- presidential_debates_2012 
dat

corp <- corpus(dat, text_field = "dialogue")