Psi Chi R - October 2024

Start Date: Oct. 7, 2024

Report Date: 31 October 2024

Source: Psi Chi R

knitr::opts_chunk$set(echo = T,message = F,warning = F)
library(tidyverse)

data=read.csv('data.csv')
#write.csv(data,'data.csv')

## EDA
SmartEDA::ExpData(data)

skimr::skim(data)

Data summary

Name	data
Number of rows	3004
Number of columns	12
_______________________
Column type frequency:
character	9
numeric	3
________________________
Group variables	None

Variable type: character

skim_variable	n_missing	complete_rate	min	max	empty	n_unique	whitespace
Title	0	1	9	62	0	2982	0
Genres	0	1	0	80	1	243	0
Release_Date	0	1	8	10	0	1252	0
Release	0	1	2	20	0	69	0
ParentalRating	0	1	0	9	1663	12	0
Plot	0	1	0	547	1	2987	0
Cast	0	1	0	308	13	2974	0
Language	0	1	0	64	69	162	0
FilmingLocations	0	1	0	125	1131	1027	0

Variable type: numeric

skim_variable	n_missing	complete_rate	mean	sd	p0	p25	p50	p75	p100	hist
Rating	232	0.92	5.06	1.46	1	4	5	6	9.8e+00	▁▆▇▃▁
RunTime	487	0.84	90.75	14.57	12	82	90	96	1.8e+02	▁▂▇▁▁
Budget	2094	0.30	3433828.09	9269568.33	1	25000	475000	2500000	8.0e+07	▇▁▁▁▁

Data processing (level 1)

### Write a script that will filter out movies that are missing a value in the ‘Budget,’ ‘RunTime,’ ‘ParentalRating’ and ‘Rating’ columns.

data1 = data %>% 
  filter(complete.cases(Budget),
         complete.cases(RunTime),
         complete.cases(ParentalRating),
         complete.cases(Rating)) %>% 
  mutate(DATE= mdy(Release_Date)) %>% 
  select(DATE,Release_Date,everything())

### Right now, the genre columns cannot be properly analyzed due to formatting (e.g., Action| Fantasy| Horror| Mystery| Romance). Write a script that will separate the column into multiple Genre categories.

data2 = data1 %>% 
  separate_wider_delim(Genres, delim = "|",
                       names_sep = "_",
                       too_few = "align_start")


# weird how Genres1 became sensible words
# data3 = data2 %>% 
#   mutate(Genres2 = strsplit(as.character(Genres1),"\\,")) %>% 
#   unnest() %>% 
#   select(Genres2,Genres1,Genres,everything())

Descriptive Statistics (level 2)

### Write a script that will provide the average, standard deviation, median, and range of the ‘Rating’ variable. Note the values you got from running the code

summary(data2$Rating,na.rm=T)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   1.000   3.900   4.900   4.955   6.000   9.600

sd(data2$Rating,na.rm = T)

## [1] 1.493

# Average = 5.0, Stand. Dev. = 1.5,
# Median = 4.9, Range = 1 to 9.6

### FILTER BY MAIN GENRE
### Write a script that will show the average ratings by a movie’s main genre. 

data2 %>% 
  group_by(Title,Genres_1) %>% 
  summarise(Rating = mean(Rating,na.rm = T)) %>% 
  arrange(desc(Rating))

Data visualization (level 3)

### Create a graph that shows the ‘Budget’ for the movies in the dataset over time.

data2 %>% 
  distinct(Title,DATE,.keep_all = T) %>% 
  ggplot(aes(x=DATE,y=Budget))+
  geom_line()+
  theme_bw()

# Something weird is happening with the Release_Date column showing 1905 for movies like Habit (2017) or Voidfinder (2017)

# extract year out of `Title` instead

data3 = data2 %>% 
  mutate(YEAR = str_extract(Title, "\\(\\d{4}\\)")) %>% 
  mutate(YEAR1 = str_extract(YEAR, "\\d{4}")) %>% 
  mutate(YEAR2 = as.Date(paste0(YEAR1,"-01-01"))) %>% 
  select(YEAR2,YEAR1,YEAR,everything())

# dollar format
dollars = function(x) {
  paste0("$", format(x,big.mark= ",",scientific=F))
}

data3 %>% 
  ggplot(aes(x=YEAR2,y=Budget))+
  geom_line()+
  theme_bw()+
  labs(x='Release Year',
       y='Movie Budget',
       title='From Box Office to Blockbuster',
       subtitle='A Decade-by-Decade Look into Movie Budgets')+
  theme(plot.title = element_text(hjust = .5),
        plot.subtitle = element_text(hjust = .5),
        axis.text.y=element_text(vjust=0,angle = 45))+
  scale_y_continuous(labels = dollars)

Inferential statistics (level 4)

### Do movies with bigger budgets get better ratings?
qqplot(data3$Budget,data3$Rating)

#No, there's no correlation between budget and rating.
cor.test(data3$Budget,data3$Rating,method = 'kendall')

## 
##  Kendall's rank correlation tau
## 
## data:  data3$Budget and data3$Rating
## z = 0.68657, p-value = 0.4924
## alternative hypothesis: true tau is not equal to 0
## sample estimates:
##        tau 
## 0.01723055

data3 %>% 
  ggplot(aes(x=Budget,y=Rating,col=Budget))+
  geom_point(size=2)+
  theme_bw()+
  scale_x_continuous(labels = dollars)

### Do movies in the Horror genre get better ratings than Mystery movies? 

#filter isn't showing anything. Odd.
#Check WS

horror= data3 %>%
  rowwise() %>%
  filter(any(c_across(starts_with("Genres_")) == " Horror"))

mystery = data3 %>%
  rowwise() %>%
  filter(any(c_across(starts_with("Genres_")) == " Mystery"))

shapiro.test(mystery$Rating)

## 
##  Shapiro-Wilk normality test
## 
## data:  mystery$Rating
## W = 0.99119, p-value = 0.7081

shapiro.test(horror$Rating)

## 
##  Shapiro-Wilk normality test
## 
## data:  horror$Rating
## W = 0.99551, p-value = 0.03097

hist(mystery$Rating)

hist(horror$Rating)

t.test(mystery$Rating,horror$Rating)

## 
##  Welch Two Sample t-test
## 
## data:  mystery$Rating and horror$Rating
## t = 3.8553, df = 153.9, p-value = 0.0001694
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  0.2547602 0.7902126
## sample estimates:
## mean of x mean of y 
##  5.477982  4.955495

#Yes, Horror does get better ratings than Mystery.