Week6
2024-10-08
data <- read.csv("C:\\Users\\gajaw\\OneDrive\\Desktop\\STATS\\vgsales.csv")Creating new variables
Variable - Deviation Total Sales
data$deviation_total_sales = abs(mean(data$Global_Sales, na.rm = TRUE) - data$Global_Sales)
head(data,15)## Rank Name Platform Year Genre Publisher
## 1 1 Wii Sports Wii 2006 Sports Nintendo
## 2 2 Super Mario Bros. NES 1985 Platform Nintendo
## 3 3 Mario Kart Wii Wii 2008 Racing Nintendo
## 4 4 Wii Sports Resort Wii 2009 Sports Nintendo
## 5 5 Pokemon Red/Pokemon Blue GB 1996 Role-Playing Nintendo
## 6 6 Tetris GB 1989 Puzzle Nintendo
## 7 7 New Super Mario Bros. DS 2006 Platform Nintendo
## 8 8 Wii Play Wii 2006 Misc Nintendo
## 9 9 New Super Mario Bros. Wii Wii 2009 Platform Nintendo
## 10 10 Duck Hunt NES 1984 Shooter Nintendo
## 11 11 Nintendogs DS 2005 Simulation Nintendo
## 12 12 Mario Kart DS DS 2005 Racing Nintendo
## 13 13 Pokemon Gold/Pokemon Silver GB 1999 Role-Playing Nintendo
## 14 14 Wii Fit Wii 2007 Sports Nintendo
## 15 15 Wii Fit Plus Wii 2009 Sports Nintendo
## NA_Sales EU_Sales JP_Sales Other_Sales Global_Sales deviation_total_sales
## 1 41.49 29.02 3.77 8.46 82.74 82.20256
## 2 29.08 3.58 6.81 0.77 40.24 39.70256
## 3 15.85 12.88 3.79 3.31 35.82 35.28256
## 4 15.75 11.01 3.28 2.96 33.00 32.46256
## 5 11.27 8.89 10.22 1.00 31.37 30.83256
## 6 23.20 2.26 4.22 0.58 30.26 29.72256
## 7 11.38 9.23 6.50 2.90 30.01 29.47256
## 8 14.03 9.20 2.93 2.85 29.02 28.48256
## 9 14.59 7.06 4.70 2.26 28.62 28.08256
## 10 26.93 0.63 0.28 0.47 28.31 27.77256
## 11 9.07 11.00 1.93 2.75 24.76 24.22256
## 12 9.81 7.57 4.13 1.92 23.42 22.88256
## 13 9.00 6.18 7.20 0.71 23.10 22.56256
## 14 8.94 8.03 3.60 2.15 22.72 22.18256
## 15 9.09 8.59 2.53 1.79 22.00 21.46256Deviation from Total Sales shows how far a game’s sales differ from the average (e.g., +10M or -5M).
Deviation from Sales identifies outliers: high-selling games or underperformers.
We can further investigate why some games have large deviations in total sales and if these outliers are due to special factors like marketing, platform popularity, or genre?
Variable - Deviation Year
data$Year <- as.numeric(as.character(data$Year))## Warning: NAs introduced by coerciondata$Year[is.na(data$Year)] <- mean(data$Year, na.rm = TRUE)
data$deviation_year = abs(mean(data$Year, na.rm = TRUE) - data$Year)
head(data, 15)## Rank Name Platform Year Genre Publisher
## 1 1 Wii Sports Wii 2006 Sports Nintendo
## 2 2 Super Mario Bros. NES 1985 Platform Nintendo
## 3 3 Mario Kart Wii Wii 2008 Racing Nintendo
## 4 4 Wii Sports Resort Wii 2009 Sports Nintendo
## 5 5 Pokemon Red/Pokemon Blue GB 1996 Role-Playing Nintendo
## 6 6 Tetris GB 1989 Puzzle Nintendo
## 7 7 New Super Mario Bros. DS 2006 Platform Nintendo
## 8 8 Wii Play Wii 2006 Misc Nintendo
## 9 9 New Super Mario Bros. Wii Wii 2009 Platform Nintendo
## 10 10 Duck Hunt NES 1984 Shooter Nintendo
## 11 11 Nintendogs DS 2005 Simulation Nintendo
## 12 12 Mario Kart DS DS 2005 Racing Nintendo
## 13 13 Pokemon Gold/Pokemon Silver GB 1999 Role-Playing Nintendo
## 14 14 Wii Fit Wii 2007 Sports Nintendo
## 15 15 Wii Fit Plus Wii 2009 Sports Nintendo
## NA_Sales EU_Sales JP_Sales Other_Sales Global_Sales deviation_total_sales
## 1 41.49 29.02 3.77 8.46 82.74 82.20256
## 2 29.08 3.58 6.81 0.77 40.24 39.70256
## 3 15.85 12.88 3.79 3.31 35.82 35.28256
## 4 15.75 11.01 3.28 2.96 33.00 32.46256
## 5 11.27 8.89 10.22 1.00 31.37 30.83256
## 6 23.20 2.26 4.22 0.58 30.26 29.72256
## 7 11.38 9.23 6.50 2.90 30.01 29.47256
## 8 14.03 9.20 2.93 2.85 29.02 28.48256
## 9 14.59 7.06 4.70 2.26 28.62 28.08256
## 10 26.93 0.63 0.28 0.47 28.31 27.77256
## 11 9.07 11.00 1.93 2.75 24.76 24.22256
## 12 9.81 7.57 4.13 1.92 23.42 22.88256
## 13 9.00 6.18 7.20 0.71 23.10 22.56256
## 14 8.94 8.03 3.60 2.15 22.72 22.18256
## 15 9.09 8.59 2.53 1.79 22.00 21.46256
## deviation_year
## 1 0.4064433
## 2 21.4064433
## 3 1.5935567
## 4 2.5935567
## 5 10.4064433
## 6 17.4064433
## 7 0.4064433
## 8 0.4064433
## 9 2.5935567
## 10 22.4064433
## 11 1.4064433
## 12 1.4064433
## 13 7.4064433
## 14 0.5935567
## 15 2.5935567Deviation from Mean Year shows how far each game’s release year is from the average release year (around 2006). This helps us understand if a game was released earlier or later compared to most other games.
Deviation from Mean Year can show trends over time. If most games have a large deviation, it means the release period for games was more spread out, while small deviations show a concentration of releases around a certain time.
We can further delve deeper to know what factors led to games being released earlier or later than the average year and if there was a technological breakthrough, a console release, or a trend that influenced game production during certain years.
Visualization
Visualization for pair 1: Global Sales and Deviation Total Sales
library(ggplot2)
ggplot(data, aes(x = deviation_total_sales, y = Global_Sales)) +
geom_point() +
geom_smooth(method = "lm", col = "blue") +
labs(title = "Global Sales vs. Deviation from Total Sales",
x = "Deviation from Total Sales (in millions)",
y = "Global Sales (in millions)")## `geom_smooth()` using formula = 'y ~ x'
The plot shows a clear linear relationship between Global Sales and Deviation from Total Sales. This indicates that as the deviation from the total sales increases, global sales also increase, following a nearly perfect positive linear trend.
Outliers- At the far right of the plot, there is a noticeable outlier with a very high Deviation from Total Sales and corresponding high Global Sales. This point represents a video game that has significantly higher sales than most other games in the dataset. Most data points are clustered around lower deviation values, indicating that most games have sales figures close to the overall mean.
Visualization for pair 2: Year and Deviation Year
#Year vs. deviation_year
ggplot(data, aes(x = deviation_year, y = Year)) +
geom_point() +
geom_smooth(method = "lm", col = "red") +
labs(title = "Year vs. Deviation from Mean Year",
x = "Deviation from Mean Year",
y = "Year")## `geom_smooth()` using formula = 'y ~ x'
In terms of potential outliers, no extreme outliers are visually evident. The data points appear to follow a relatively consistent pattern along the negative slope, with some mild deviations from the line.
There is a negative linear trend in this plot, which is expected. The higher the deviation from the mean year, the further the data points are from the mean release year (which appears to be around 2005–2010).
- This is a logical outcome because as the year moves away from the mean year (both in the past and future), the deviation increases symmetrically.
However, there is a slightly wider spread of points between 2010 and 2020. Some recent game release years (those after 2010) are scattered further from the regression line, which could indicate some irregularity in the release patterns in more recent years.
Similarly, there are a few points before 1985 that could be considered outliers compared to the rest of the data. These might represent very early video games that were rare or released at times when the video game industry was not as developed.
Calculating Correlation Coefficient
Correlation between Global Sales and Deviation from Total Sales
cor(data$Global_Sales, data$deviation_total_sales)## [1] 0.9675341This strong positive correlation makes perfect sense based on the plot. It indicates a nearly perfect linear relationship, meaning that as the deviation from total sales increases, global sales also increase almost proportionally. This high correlation reflects that games with larger sales deviations (especially blockbusters) dominate global sales numbers.
Correlation between Year and Deviation from Mean Year
cor(data$Year, data$deviation_year)## [1] -0.3817814The moderate negative correlation is consistent with the plot, showing a general downward trend but with some variability. While the relationship between Year and Deviation from Mean Year is negative, it’s not as strong due to the scattering of data points, particularly for recent years. This indicates that the deviation from the mean year doesn’t fully explain the variation in release years but still shows a meaningful trend.
Calculating Confidence Intervals
Confidence Interval for Global Sales:
#mean and standard error for Global Sales
mean_global_sales <- mean(data$Global_Sales, na.rm = TRUE)
std_error_global_sales <- sd(data$Global_Sales, na.rm = TRUE) / sqrt(nrow(data))
# Calculating the 95% confidence interval for Global Sales
ci_lower_global_sales <- mean_global_sales - qt(0.975, df=nrow(data)-1) * std_error_global_sales
ci_upper_global_sales <- mean_global_sales + qt(0.975, df=nrow(data)-1) * std_error_global_sales
#result
ci_lower_global_sales## [1] 0.513782ci_upper_global_sales## [1] 0.5610993Insight:
- The 95% confidence interval for Global Sales is between 0.514 million and 0.561 million units.
This means we are 95% confident that the average global sales of all games in the population fall between these values.
The interval is quite narrow, indicating that the variation in global sales is relatively small for most games in this dataset. Most games tend to have sales that are quite close to the average.
Significance:
- A narrow confidence interval suggests that the sample we have is reliable and gives a precise estimate of the true average sales for all games.
- This level of confidence can be useful for predicting expected sales of future games, especially for developers and publishers who want to benchmark their game’s performance against past successes.
Further Questions :
Why is the average global sales relatively low? Does this mean that a few blockbuster games (outliers) are skewing perceptions of overall success, while most games sell closer to the lower end of the spectrum?
How do regional sales compare? Are there regions (e.g., North America, Europe, Japan) where games are more likely to sell above or below this global average?
Does the confidence interval change significantly over different periods of time? For example, do older games (e.g., from the 1980s or 1990s) have a different sales pattern than more recent ones?
Confidence Interval for Year:
#mean and standard error for year
mean_year <- mean(data$Year, na.rm = TRUE)
std_error_year <- sd(data$Year, na.rm = TRUE) / sqrt(nrow(data))
# Calculating the 95% confidence interval for Year
ci_lower_year <- mean_year - qt(0.975, df=nrow(data)-1) * std_error_year
ci_upper_year <- mean_year + qt(0.975, df=nrow(data)-1) * std_error_year
#result
ci_lower_year## [1] 2006.318ci_upper_year## [1] 2006.494Insights:
The 95% confidence interval for the release year is between 2006.318 and 2006.494.
This means that we are 95% confident that the average release year of games in the dataset falls between 2006 and 2007.
The interval is very narrow, indicating that most of the games in this dataset were released around that time, suggesting a concentration of game releases in the mid-2000s.
Significance:
The narrow interval shows that the sample of games in this dataset is mostly centered around 2006, meaning that this period might represent a peak in video game releases or popularity.
This period could be significant for the gaming industry, potentially marking a time when many major titles were released, possibly due to advancements in gaming technology or the rise of popular consoles.
Further Questions:
Why is the average release year around 2006? Was there a particular event or trend in the gaming industry that caused a surge in game releases during this time?
How do game sales trends differ before and after this period? Is 2006 a turning point in terms of how successful games were, or how the industry evolved?
What consoles or platforms were popular during this time? Do certain platforms (like the Wii, Xbox, or PlayStation) dominate the sales for games released in this period?