Scarlett Johansson movies
Jonathan Bouchet
2017-04-21
options(width=100)
knitr::opts_chunk$set(out.width='1000px',dpi=200,message=FALSE,warning=FALSE)#load packages and csv file
library(ggplot2)
library(dplyr)
library(gridExtra)
library(grid)
library(ggthemes)
library(knitr)
library(ggdendro)
library(factoextra)
library(NbClust)Following 538’s analysis on the Four types of Scarlett Johansson, I had an updated look at this data (box office vs. RottenTomatoes score) in light of the release of her new movie : Ghost in the Shell
This kernel simply :
- aims to reproduce 538’s article plot
- find the new category of
GITS
Data are taken from the same source of the article ; for GITS box office and RottenTomatoes (Google)
df<-read.csv('scarlett_johansson_3.csv',sep=',')
head(df)## X date Movie Role Domestic International
## 1 1 3/31/17 Ghost in the Shell The Major 33304297 96942362
## 2 2 12/21/16 Sing Ash 269943555 343600000
## 3 3 5/6/16 Captain America: Civil War Natasha Romanoff / Black Widow 408084349 743600000
## 4 4 4/15/16 The Jungle Book Kaa 364001123 599900000
## 5 5 2/5/16 Hail, Caesar! DeeAnna Moran 30080225 34091194
## 6 6 5/1/15 Avengers: Age of Ultron Natasha Romanoff/Black Widow 459005868 945700000
## Worldwide RottenTomatoes CAT DATE year
## 1 130246659 46 OTHER 2017-03-31 2017
## 2 613543555 73 AVENGERS 2016-12-21 2016
## 3 1151684349 90 AVENGERS 2016-05-06 2016
## 4 963901123 95 AVENGERS 2016-04-15 2016
## 5 64171419 86 HER 2016-02-05 2016
## 6 1404705868 75 AVENGERS 2015-05-01 20151 Plots as 538
The categories define here as the same as W. Hickey in his article. Even by eye we can see that GITS may fit in the Peers category : low RottenTomatoes score and low revenues at the box office.
g1<-ggplot() +
geom_point(data=filter(df,CAT=='AVENGERS'),aes(x=RottenTomatoes,y=Domestic,color='Avengers'),size=3) +
geom_point(data=filter(df,CAT=='HER'),aes(x=RottenTomatoes,y=Domestic,color='Her'),size=3) +
geom_point(data=filter(df,CAT=='PEERS'),aes(x=RottenTomatoes,y=Domestic,color='Peers'),size=3) +
geom_point(data=filter(df,CAT=='PRESTIGIOUS'),aes(x=RottenTomatoes,y=Domestic,color='Prestigious'),size=3) +
geom_point(data=filter(df,CAT=='OTHER'),aes(x=RottenTomatoes,y=Domestic,color='GITS'),size=3) +
scale_color_manual(name="",values = c(Avengers="orange",Her="#46ACC8",Peers="pink",Prestigious="#0B775E",GITS="#F21A00")) + theme_fivethirtyeight() + xlim(0,100) + theme(legend.position="top") + xlab('Rotten Tomatoes') + ylab('Domestic box office') + ggtitle("Domestic box office ($US) vs. RottenTomatoes score")
g1
Taking the yearly average of the RottenTomatoes score of the movies reeased in a given year, we can see when Johansson was hype or less hype.
g2<-df %>%
group_by(year) %>%
select(year,RottenTomatoes) %>%
summarise(meanRT = mean(RottenTomatoes)) %>%
ggplot(aes(x=year,y=meanRT)) +
geom_line(size=1.5,color='#E2D200') +
geom_point(aes(color=ifelse(year==2017,'gits','old')),size=2) +
ylim(0,100) +
scale_color_manual(name="",values = c(gits="#F21A00",old="black")) +
theme_fivethirtyeight() +
theme(legend.position='none') + xlab('year') + ylab('average Rotten Tomatoes') +
ggtitle("RottenTomatoes movies score vs. Year")
g2
#grid.arrange(g1,g2,ncol=2)2 Finding clusters
The goal here was to check/find new clusters ##Dendogram
tt <-data.frame(df %>% select(Movie,Domestic,RottenTomatoes))
rownames(tt)<-tt$Movie
tt$Movie<-NULL
dd <- dist(scale(tt), method = "euclidean")
hc <- hclust(dd, method = "ward.D2")
ggdendrogram(hc, rotate = FALSE, theme_dendro = FALSE) + xlab('') + ylab('')
2.1 Optimal cut
The optimal cut is based on the Within cluster sum of squares to find meaningful clusters.
nb <- NbClust(scale(tt), distance = "euclidean", min.nc = 2,max.nc = 10, method = "ward.D2", index ="all")
## *** : The Hubert index is a graphical method of determining the number of clusters.
## In the plot of Hubert index, we seek a significant knee that corresponds to a
## significant increase of the value of the measure i.e the significant peak in Hubert
## index second differences plot.
## 
## *** : The D index is a graphical method of determining the number of clusters.
## In the plot of D index, we seek a significant knee (the significant peak in Dindex
## second differences plot) that corresponds to a significant increase of the value of
## the measure.
##
## *******************************************************************
## * Among all indices:
## * 2 proposed 2 as the best number of clusters
## * 12 proposed 3 as the best number of clusters
## * 5 proposed 4 as the best number of clusters
## * 1 proposed 7 as the best number of clusters
## * 3 proposed 10 as the best number of clusters
##
## ***** Conclusion *****
##
## * According to the majority rule, the best number of clusters is 3
##
##
## *******************************************************************fviz_nbclust(nb)## Among all indices:
## ===================
## * 2 proposed 0 as the best number of clusters
## * 1 proposed 1 as the best number of clusters
## * 2 proposed 2 as the best number of clusters
## * 12 proposed 3 as the best number of clusters
## * 5 proposed 4 as the best number of clusters
## * 1 proposed 7 as the best number of clusters
## * 3 proposed 10 as the best number of clusters
##
## Conclusion
## =========================
## * According to the majority rule, the best number of clusters is 3 .
grp<-cutree(hc, k=3)
fviz_cluster(list(data = tt, cluster = grp)) + ggtitle('')
tt$group<-grp
ggplot() + geom_point(data=tt,aes(x=RottenTomatoes,y=Domestic,color=factor(group)),size=3) +
scale_color_manual(name="",values = c("#46ACC8","#0B775E","#F21A00")) +
theme_fivethirtyeight() + xlim(0,100) +
theme(legend.position="top") + xlab('Rotten Tomatoes') + ylab('Domestic box office') + ggtitle("Domestic box office ($US) vs. RottenTomatoes score")
NbClustwasn’t able to differentiateHerandPrestigiousclusters from the 538 analysis using hte optimal number of clusters,- as expected,
GITSmay be categorized in thePeerscategory
History :
- version 1 : initial commit