TedTalks

library(stringr)
library(tm)
library(wordcloud)
library(wesanderson)
library(ggplot2)
library(dplyr)
library(gridExtra)
library(readxl)
library(tidyr)
ted <- read_xlsx("/Users/Bryce/DataViz/TedTalks/TED Talks as of 3.29.2018.xlsx")
colnames(ted)

##  [1] "Talk ID"      "public_url"   "speaker_name" "headline"     "description" 
##  [6] "event"        "duration"     "language"     "published"    "tags"

Clean the Data

clean.text <- function(x) {
    # to lowercase
    x <- tolower(x)
    # remove punctuation marks
    x <- gsub("[[:punct:]]", "", x)
    # remove numbers
    x <- gsub("[[:digit:]]", "", x)
    # remove tabs and extra spaces
    x <- gsub("[ |\t]{2,}", "", x)
    # remove blank spaces at the beginning
    x <- gsub("^ ", "", x)
    # remove blank spaces at the end
    x <- gsub(" $", "", x)
    # result
    return(x)
}

View Number of Speakers by Number of TED Talks

numtalks <- data.frame(table(ted$speaker_name))
table(numtalks$Freq)

## 
##    1    2    3    4    5    6    7    9 
## 2047  208   51   18    6    2    1    1

2047 speakers have only done one TED Talk whereas one speaker has done a record nine TED Talks

Get TED Talk Time in Minutes

# Remove the 1899 Date from each duration entry
ted$duration <- gsub('1899-12-31 ', '', ted$duration)
# Create function to display talk time in minutes
ted$TalkTime <- sapply(strsplit(as.character(ted$duration), ":"), function(x) {
    x <- as.numeric(x)
    x[1] * 60 + x[2] + x[3]/60
})

Establish Mean Time Talked in Frequency and Duration Table

speakfreqandduration <- ted %>% group_by(speaker_name) %>% summarise(NumTalks = n(), 
    Mean.Talk.Time = mean(TalkTime, na.rm = TRUE)) %>% filter(NumTalks > 
    2)
# View talk time stats
summary(speakfreqandduration$Mean.Talk.Time)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   3.033  11.081  14.581  14.495  17.286  36.339

Interactive Table with Speakers, Number of Talks, and Mean Talk Time

library(htmlwidgets)
suppressPackageStartupMessages(library(dplyr))
library(DT)# if (!require("DT")) install.packages('DT')
dt=datatable(speakfreqandduration, options=list(pageLength = 15), colnames = c('Speaker Name','Number of Talks','Average Talk Time (min)'))
dt

Mean Time of Speakers Who Gave More Than 3 Talks

gg3 <- ggplot(na.omit(speakfreqandduration), aes(x = reorder(speaker_name, Mean.Talk.Time), 
    y = Mean.Talk.Time, fill = as.factor(NumTalks))) + geom_bar(stat = "identity") + 
    xlab("Speaker") + theme_bw() + theme(axis.text.y = element_text(size = 8), plot.title = element_text(hjust = 0.5), 
    axis.title.y = element_blank()) + coord_flip() + ggtitle("Speakers, Mean Talk Time, and Number of Talks") +guides(fill=guide_legend(title="Number of TED Talks Given")) + ylab("Mean Talk Time (Minutes)")

gg3

Get Top 20 Tags Using Wordmap Functions and Plot

#Delineated with spaces instead of commas
tagCorpus <- Corpus(VectorSource(gsub(","," ",ted$tags)))
tagStopwords <- c(stopwords("english"), "the")


tagCorpus <- tm_map(tagCorpus, removeWords, tagStopwords)

## Warning in tm_map.SimpleCorpus(tagCorpus, removeWords, tagStopwords):
## transformation drops documents

tdmpremat_tag <- TermDocumentMatrix(tagCorpus)
tdm_tag <- as.matrix(tdmpremat_tag)
sortedMatrix_tag <- sort(rowSums(tdm_tag), decreasing = TRUE)
tdmframe_tag <- data.frame(word = names(sortedMatrix_tag), freq = sortedMatrix_tag)

top_20_tags <- top_n(tdmframe_tag,20)

## Selecting by freq

ggplot(top_20_tags, aes(reorder(word, freq), freq)) + geom_bar(stat='identity') + coord_flip() + xlab("Tag") + ylab("Frequency") + ggtitle("Most Common Video Tags") + theme(plot.title = element_text(hjust = 0.5))

Extra Non-assigned Graphs and Figures

Filter Data and Plot Word Cloud of Common Words in TED Talk Titles

myCorpus <- Corpus(VectorSource(clean.text(ted$headline)))
myStopwords <- c(stopwords("english"), "ted prize wish")


myCorpus <- tm_map(myCorpus, removeWords, myStopwords)
tdmpremat <- TermDocumentMatrix(myCorpus)
tdm <- as.matrix(tdmpremat)
sortedMatrix <- sort(rowSums(tdm), decreasing = TRUE)
tdmframe <- data.frame(word = names(sortedMatrix), freq = sortedMatrix)


# plot words appearing >5 times
wordcloud(tdmframe$word, tdmframe$freq, random.order = FALSE, random.color = FALSE, 
    min.freq = 5, scale = c(5, 0.2), colors = wes_palette("Darjeeling1"))

Graph of Speaker Frequency who Gave 3 or More TED Talks

# Not required for assignment
gg1 <- ggplot(speakfreqandduration, aes(x = NumTalks, fill = as.factor(NumTalks))) + geom_histogram(binwidth=1) + xlab("Number of talks") + ggtitle("Frequency of Speakers Who Have Given 3 or More Talks") + theme_bw() + theme(legend.position = "none", plot.title = element_text(hjust = 0.5)) + scale_x_continuous(breaks = 1:10)
gg1