RPubs link information

• Rpubs link comes here: https://rpubs.com/Planpoint/s3957789

Introduction

• This study includes the association between game and sales performance, exploring if the genres significantly influences the success of games in the market of NA(North America), EU(Europe), JP(Japan), Other(Rest of the world), and Global sales.
  
  
  
  

Introduction Cont.

• This study will seek evidence to support the hypothesis that ‘Genre’ significantly influences sales performance.

• We will explore relationship, valuable insights to inform genre selection, marketing strategies, and overall market trends in the gaming industry.

• There are total 12 genres and 5 sale regions

• Genre is based on game-play mechanics, themes and player experiences.

Problem Statement

• To question the relationship between game genre and sales performance and how does the genre influences the success of games in the market.

• Statistical analysis will be used to examine the relationship between game genre and sales performance.

• Descriptive statistics will provide more insights into sales trends across genres, while the chi-square test of association will determine if there is a significant dependency.

• Inferential statistics will quantify the strength of the relationship and inform recommendations for genre selection and sales strategies.

Data

• The data required was collected from this website: https://www.kaggle.com/code/snanilim/video-games-sales-analysis-and-visualization/notebook

• Variables:

  • Rank[type int] :- Overall sales ranking
  • Name[type chr] :- Name of games
  • Platform[type obj] :- Game release platforms
  • Year[type float] :- Year of the game release
  • Genre[type obj] :- Game genre
  • Publisher[type obj] :- Game publisher
  • NA_Sales[type float] :- Statistical Sales data in North America(in millions)
  • EU_Sales[type float] :- Statistical Sales data in Europe(in millions)
  • JP_Sales[type float] :- Statistical Sales data in Japan(in millions)
  • Other_Sales[type float] :- Statistical Sales data in the rest of the world(in millions)
  • Global_Sales[type float] :- Total Statistical Sales worldwide

Imported Data

game_sales <- read.csv('vgsales.csv')
knitr::kable(head(game_sales,5))

Data Cleaning

• For this project proposal, the required column for the statistical analysis would be the columns Genre, NA_Sales, EU_Sales, JP_Sales, Other_Sales and Global_Sales.
• The type of all the sales column in the format of num which are in the range of 82.74 as maximum value and 0.00 as minimum value.
• Among the above mentioned columns the Genre column is an categorical column, hence it has been converted into factor format. This factor variable has 12 different levels which can be stated as :-

Data After Cleaning

knitr::kable(head(df,5))

Descriptive Statistics

df_description <- group_by(df, Genre) %>% summarise_at(vars('NA_Sales', 'EU_Sales', 'JP_Sales', 'Other_Sales', 'Global_Sales'),
                        list(Mean = mean, 
                             Median = median, 
                             Standard_Deviation=sd, 
                             Minimum = min, 
                             Maximum = max, 
                             Interquartile_Range=IQR, 
                             Quantile_1 = ~quantile(.,probs=0.25),
                             Quantile_3 = ~quantile(.,probs=0.75)))
datalong <- gather(df_description, 'Variable', 'Value', c('NA_Sales_Mean':'Global_Sales_Quantile_3'), factor_key = TRUE)
final <- spread(datalong, Genre, Value)
final$'Summary Statistics' <- sub(".*_","", final$Variable)
final$Variable <- sub("_.*","", final$Variable)
final$`Summary Statistics`[final$`Summary Statistics` == '1'] <- 'Quartile 1'
final$`Summary Statistics`[final$`Summary Statistics` == '3'] <- 'Quartile 3'

temp_final <- final
temp_final <- temp_final[,c(1, 14, 2,3,4,5,6,7,8,9,10,11,12,13)]
counter=1
for (i in 1:nrow(final)){
  if (i %in% c(1,6,11,16,21,26,31,36,41)){
    next
  }
  else{
    temp_final$`Summary Statistics`[i] <- ''
  }
}
knitr::kable(temp_final)

Decsriptive Statistics Cont.

ggplot(df, aes(x = Genre, y = NA_Sales)) +
  geom_boxplot() +
  xlab("Genre") +
  ylab("North America Sales") +
  ggtitle("Distribution of Sale data of Different Genres in North America") 

ggplot(df, aes(x = Genre, y = EU_Sales)) +
  geom_boxplot() +
  xlab("Genre") +
  ylab("Europe Sales") +
  ggtitle("Distribution of Sale data of Different Genres in Europe")

ggplot(df, aes(x = Genre, y = JP_Sales)) +
  geom_boxplot() +
  xlab("Genre") +
  ylab("Japan Sales") +
  ggtitle("Distribution of Sale data of Different Genres in Japan")

ggplot(df, aes(x = Genre, y = Other_Sales)) +
  geom_boxplot() +
  xlab("Genre") +
  ylab("Other Sales") +
  ggtitle("Distribution of Sale data of Different Genres in Other Regions")

ggplot(df, aes(x = Genre, y = Global_Sales)) +
  geom_boxplot() +
  xlab("Genre") +
  ylab("Global Sales") +
  ggtitle("Distribution of Sale data of Different Genres at Global Level")

custom_colors <- c("#FF0000", "#00FF00", "#0000FF", "#FFFF00", "#00FFFF", "#FF00FF", "#800000", "#008000", "#000080", 
                   "#808000", "#800080", "#008080", "#808080")

# Create a mosaic plot with custom colors
ggplot(df, aes(x = "", fill = Genre)) +
  geom_bar(width = 1, position = "fill") +
  coord_flip() +
  scale_fill_manual(values = custom_colors) +
  xlab("") +
  ylab("Proportion") +
  ggtitle("Sale Distribution With Respect To Genre")

Hypothesis Testing

• Hypothesis:

  • Null hypothesis (H0): There is no association between game genre and sales performance. \[H_0: \mu_1 = \mu_2 \]

  • Alternative hypothesis (Ha): There is an association between game genre and sales performance. \[H_A: \mu_1 \ne \mu_2\]

chiSquare_func <- function(input_table, col_name){
  # Perform chi-square test of association with expected frequencies
  result <- chisq.test(input_table)
  
  # Calculate expected frequencies
  expected <- result$expected
  
  # Calculate chi-square test statistic
  chi_square <- sum((input_table - expected)^2 / expected)
  
  # Calculate degrees of freedom
  df_stat <- (nrow(input_table) - 1) * (ncol(input_table) - 1)
  
  # Calculate p-value
  p_value <- 1 - pchisq(chi_square, df_stat)
  
  # Print the result
  cat("The Chi-Square Analysis for :- ", col_name, '\n')
  cat("Chi-square test statistic:", chi_square, "\n")
  cat("Degrees of freedom:", df_stat, "\n")
  cat("p-value:", p_value, "\n")
}

table_data <- table(df$Genre, cut(df$NA_Sales, breaks = c(-Inf, 5, 10, 15, Inf), labels = c("Low", "Medium", "High", "Very High")))
chiSquare_func(table_data, "NA_Sales")

## The Chi-Square Analysis for :-  NA_Sales 
## Chi-square test statistic: 102.2899 
## Degrees of freedom: 33 
## p-value: 5.142919e-09

table_data <- table(df$Genre, cut(df$EU_Sales, breaks = c(-Inf, 3, 5, 8, Inf), labels = c("Low", "Medium", "High", "Very High")))
chiSquare_func(table_data, 'EU_Sales')

## The Chi-Square Analysis for :-  EU_Sales 
## Chi-square test statistic: 49.55325 
## Degrees of freedom: 33 
## p-value: 0.03212222

table_data <- table(df$Genre, cut(df$JP_Sales, breaks = c(-Inf, 0.5, 1, 1.5, Inf), labels = c("Low", "Medium", "High", "Very High")))
chiSquare_func(table_data, 'JP_Sales')

## The Chi-Square Analysis for :-  JP_Sales 
## Chi-square test statistic: 401.2928 
## Degrees of freedom: 33 
## p-value: 0

table_data <- table(df$Genre, cut(df$Other_Sales, breaks = c(-Inf, 1, 1.5, 2.5, Inf), labels = c("Low", "Medium", "High", "Very High")))
chiSquare_func(table_data, 'Other_Sales')

## The Chi-Square Analysis for :-  Other_Sales 
## Chi-square test statistic: 42.72626 
## Degrees of freedom: 33 
## p-value: 0.1196447

table_data <- table(df$Genre, cut(df$Global_Sales, breaks = c(-Inf, 10, 20, 30, Inf), labels = c("Low", "Medium", "High", "Very High")))
chiSquare_func(table_data, 'Global__Sales')

## The Chi-Square Analysis for :-  Global__Sales 
## Chi-square test statistic: 68.15063 
## Degrees of freedom: 33 
## p-value: 0.0003062696

Hypthesis Testing Cont.

• The Formulae used for making the above Chi-Square predictions are as below:
  • Chi-Square Statistic :-

\[\chi^2 = \sum \frac{{(O_{ij} - E_{ij})^2}}{{E_{ij}}} \] + Degree of Freedom :- \[df = (r - 1) * (c - 1)\]

Discussion

• CASE 1:-
  • The p-value associated with the chi-square tests are 5.142919e-09, 0.03212222, 0.0003062696 for NA_Sales, EU_Sales and Global_Sales.
  • The p-value is a measure of the statistical evidence against the null hypothesis.
  • A p-value below a chosen significance level (e.g., 0.05) indicates that there is strong evidence to reject the null hypothesis and we conclude that there is a significant association between the variables.
  • This concludes that there is strong relation between genre of the games and sale of that game in the given regions.
• CASE 2:-
  • The p-value associated with the chi-square test is 0 for that of JP_Sales.
  • A p-value of 0 indicates that the observed data is extremely unlikely to occur under the assumption of the null hypothesis.
  • It suggests strong evidence to reject the null hypothesis and conclude that there is a significant association between the variable JP_Sales and other variables.
• CASE 3:-
  • The p-value associated with the chi-square test is 0.1196447.
  • The p-value is a measure of the statistical evidence against the null hypothesis.
  • A p-value greater than the chosen significance level (e.g., 0.05) suggests that the observed data is not statistically significant and does not provide strong evidence to reject the null hypothesis of no association between the variable Other_Sales and other variables.
  • This implies that sampling data for Japan is not adequate enough to infer the insights mentioned in the report.
• In Summary, the Chi-square analysis gives an in-depth result for each and every region considered in the data set except for Japan.
• This Resulted in Genre being important for sale in Japan, North America, Europe and Globally, deducting other regions.

References

• Diana Mindrila, Phoebe Balentyne, The Chi Square Test, accessed on 28th May, 2023. chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/https://www.westga.edu/academics/research/vrc/assets/docs/ChiSquareTest_LectureNotes.pdf
• Syed Abdul Hadi, Chi Square Test in R, Accessed on 29th May, 2023. https://www.programmingr.com/statistics/chi-square-test/
• Yihui Xie, J. J. Allaire, Garrett Grolemund, R Markdown: The Definitive Guide, accessed on 31st May, 2023. https://bookdown.org/yihui/rmarkdown/
• GREGORYSMITH, Video game Sales, accessed on 25th May, 2023 https://www.kaggle.com/code/snanilim/video-games-sales-analysis-and-visualization/notebook