Longecity Forum Nicotinamide Riboside [curated] Thread Post Ratings

library(data.table)
library(dplyr)
library(ggplot2)
library(lubridate)
library(RColorBrewer)
library(rvest)

urls <- paste0( 'http://www.longecity.org/forum',
                '/topic/82770-nicotinamide-riboside-curated')
pages <-
  read_html(urls) %>%
  html_node(xpath='(//*[@class="last"]/a)[1]/@href') %>%
  html_text() %>%
  gsub(pattern='.*-', replacement='')
if (!is.na(pages)) urls <- paste0(urls, paste0('/page-', seq_len(pages)))

ratings <- 
  do.call( rbind, lapply( urls, function(url) {
    posts <- html_nodes( read_html(url), '.post_body')
    do.call( rbind, lapply( posts, function(post) {
      ratings <- post %>%
        html_nodes('.ratingtitle') %>%
        html_text()
      if (length(ratings)) {
        date <- post %>%
          html_nodes('.published') %>%
          html_attr('title') %>%
          as_date()
        count <- post %>%
          html_nodes('.ratingfreq') %>%
          html_text() %>%
          as.numeric()
        data.table( date, rating = ratings, frequency = count)
      }
    }))
  }))

rating_types <- c(
  'Good Point' = "Positive",
  'Informative' = "Positive",
  'Cheerful' = "Positive",
  'Well Written' = "Positive",
  'WellResearched' = "Positive",
  'like' = "Neutral",
  'Agree' = "Neutral",
  'Enjoying the show' = "Neutral",
  'Needs references' = "Neutral",
  'unsure' = "Neutral",
  'Disagree' = "Neutral",
  'dislike' = "Neutral",
  'Ill informed' = "Negative",
  'Off-Topic' = "Negative",
  'Dangerous, Irresponsible' = "Negative",
  'Pointless, Timewasting' = "Negative",
  'Unfriendly' = "Negative")
ratings[, period := format(date, "%Y-%m")]

ratings[, rating := factor(rating, levels = names(rating_types))]

ratings[, type := factor(rating_types[rating], levels=unique(rating_types))]

rating_colors <- c( rev(brewer.pal(6,"Greens")[-1]),
                    rev(brewer.pal(4, "Blues")[-1]),
                    "#FFFFFF",
                    brewer.pal(4, "Purples")[-1],
                    brewer.pal(6,"Reds")[-1])
names(rating_colors) <- names(rating_types)
type_colors <- c('Negative' = '#FB6A4A',
                 'Neutral' = "#C5CFE4",
                 'Positive' = '#74C476')

Total rating percentages.

ratings %>%
  group_by(rating) %>%
  tally(frequency) %>%
  mutate(n = round(100 * n / sum(n))) %>%
  arrange(rating) %>%
  ggplot(data = ., aes(x = factor(1), y = n, fill = rating)) +
    geom_bar(stat = 'identity', width = 1) +
    scale_fill_manual(values = rating_colors) +
    coord_polar(theta = "y") +
    theme_void()

Rating type volume each month.

ratings %>%
  arrange(period, type) %>%
  ggplot(data = ., aes(period, frequency, fill = type)) +
    geom_bar(position = "stack", stat = "identity") +
    scale_fill_manual(values = type_colors) +
    guides(fill = guide_legend(reverse=TRUE, title = 'Rating Type')) +
    theme(axis.text.x  = element_text(angle=50, vjust = .5))

Ratings as a percent of total ratings each month.

ratings %>%
  group_by(period) %>%
  mutate(freqTtl = sum(frequency)) %>%
  group_by(period, rating) %>%
  mutate(percent = round(100 * frequency / freqTtl, 2)) %>%
  arrange(period, factor(rating, levels=names(rating_types))) %>%
  ggplot(aes(period, percent, fill = factor(rating, levels = names(rating_types)))) +
    geom_bar(position = "stack", stat = "identity") +
    scale_fill_manual(values = rating_colors) +
    guides(fill = guide_legend(reverse=TRUE, title = 'Rating')) +
    theme(axis.text.x  = element_text(angle=50, vjust = .5))