Skip to content

Instantly share code, notes, and snippets.

@pssguy
Created May 1, 2013 21:14
Show Gist options
  • Save pssguy/5498431 to your computer and use it in GitHub Desktop.
Save pssguy/5498431 to your computer and use it in GitHub Desktop.
Shiny App showing ranking of 140+ TV Shows by episode
# load required packages
library(shiny)
library(shinyIncubator)
library(googleVis)
library(ggplot2)
library(stringr)
library(plyr)
library(XML)
library(httr)
library(Hmisc)
library(changepoint)
# load pre-compiled list of shows on GEOS including title and url code
allShows<- read.csv("http://dl.dropboxusercontent.com/u/25945599/Shows.csv",stringsAsFactors=FALSE)
allShows <- subset(allShows,!is.na(title))
# set data for selectInput in
showSelection <- allShows$title
shinyServer(function(input, output) {
# main data scraping and processing
Data <-reactive( {
# scrape selected show and create data.frame
showID <- allShows[allShows$title==input$show,]$id
showURL <- paste0("http://www.geos.tv/index.php/list?sid=",showID,"&collection=all")
z <- readHTMLTable(showURL, stringsAsFactors = FALSE)
episodes <- z[["collectionTable"]]
# perform some simple tidying
episodes$Mean <- as.numeric(str_sub(episodes$Mean,1,4))
episodes$Count <- as.integer(episodes$Count)
episodes$epOrder <- as.integer(episodes[[1]])
episodes <- arrange(episodes,epOrder)
# create the changepoint object and the lines.df
# necessary for plot lines
goodData <- subset(episodes,!is.na(Mean))
pelt <- cpt.mean( goodData$Mean,method='PELT')
cpts <- pelt@cpts
st <- c(0,cpts[-length(cpts)])
means <- pelt@param.est$mean
lines.df <- data.frame(st=st,fin=cpts,means=means)
# make data.frames available to other functions
info <- list(lines.df=lines.df,episodes=episodes)
return(info)
})
# enable paging on gvisTable
myOptions <- reactive({
list(
page='enable',
pageSize=15
)
})
# Use gvisTable to enable paging and sorting
output$gvisTable <- renderGvis( {
# make more presentable
df <- Data()$episodes[,c("epOrder","Title","Mean","Count")]
names(df) <- c("Episode","Title","Av Rating","Rankers")
df <- subset(df,Rankers>0&Rankers!="")
gvisTable(df, options=myOptions())
})
output$plot <- renderPlot( {
plotdf <- Data()$episodes
linesdf <- Data()$lines.df
maxCount <- max(subset(plotdf,Count>0)$Count)
print(
ggplot(subset(plotdf,Count>0), aes(x=epOrder,y=Mean))+geom_point(alpha=subset(plotdf,Count>0)$Count/maxCount)+
geom_segment(data=linesdf, aes(x = st, y = means, xend = fin, yend = means, colour="red"))+
theme(legend.position="none") +
ylab("Average Rating (out of 10)")+xlab("Episode Order")
)
})
output$notes <- renderUI( {
df <-subset(Data()$episodes,!is.na(Count))
max <- max(df$Count)
min <- min(df$Count)
mean <-ceiling(mean(df$Count))
HTML(paste0("The graph represents the average ranking for the show over time. The red lines
indicate changepoints, estimations of when the properties of the time-series, typically the mean changes.
The intensity of the plot varies according to the number of respondents. An episode of a show
that is favourably rated tends to get more people ranking as do earlier episodes in long-running show.<p><p> For ",input$show," the average number of rankers was
",mean," with a maximum of ",max))
})
})
shinyUI(pageWithSidebar(
# Application title
headerPanel("TV Show Rankings"),
# Sidebar with information, controls to select the player and a best-of table
sidebarPanel(
helpText(
p("Choose from one of 145 popular TV shows to see episode ranking by ",a("GEOS", href="http://www.geos.tv/")," members.")
),
wellPanel(
selectInput("show", "Select Show:",showSelection)
),
p("Regular Articles - ",
a("PSS blog", href="http://premiersoccerstats.com/wordpress/")
),
p("Twitter Feed - ",
a("@pssguy", href="http://twitter.com/pssGuy")
),
p("Contact - ",
a( "andy@premiersoccerstats.com", href="mailto:andy@premiersoccerstats.com"))
),
mainPanel(
tabsetPanel(
tabPanel("Chart", plotOutput("plot"),htmlOutput("notes")), # no good , height="200px"
tabPanel("Sortable Table", htmlOutput("gvisTable"))
)
)
))
@mattmalin
Copy link

This is really great, a particularly nice use of Shiny!

Some of the scraped titles in the csv have cut off prematurely though for shows including a ":" in the show title. e.g. Terminator: The Sarah Connor Chronicles and all the Star Trek: and Star Wars: shows, so they can't be distinguished in the list, and encoding problems mean Carnivàle also doesn't display properly.

@BoisterousCoder
Copy link

What is Shiny exactly? I have never heard of it.

@pssguy
Copy link
Author

pssguy commented Jun 21, 2013

@mattmalin. Thanks. Yes there are 1 or 2 aspects I should follow up

@pssguy
Copy link
Author

pssguy commented Jun 21, 2013

@shadowace112 From http://www.rstudio.com/shiny/

Shiny makes it super simple for R users like you to turn analyses into interactive web applications that anyone can use. Let your users choose input parameters using friendly controls like sliders, drop-downs, and text fields. Easily incorporate any number of outputs like plots, tables, and summaries.

No HTML or JavaScript knowledge is necessary. If you have some experience with R, you’re just minutes away from combining the statistical power of R with the simplicity of a web page

@xiaodaigh
Copy link

Awesomeness incarnate!

@rkillick
Copy link

rkillick commented Oct 7, 2016

Sadly the GOES website is no longer live. This was a great example that everyone could relate to. It appears as though someone has something similar but does the fits by season here: http://graphtv.kevinformatics.com/ interesting to know if we could scrap the raw data and update the shiny app to use the new data instead.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment