Video Game Sales Dataset (1980–2024)
This analysis explores sales trends and critic scores across thousands of video games spanning four decades.
Data Source: Kaggle - Video Game Sales 1980 – 2024
Key Variables:
Here’s how the data is loaded and filtered:
library(ggplot2)
library(plotly)
library(dplyr)
df <- read.csv("Video_Games_Sales_Cleaned.csv",
stringsAsFactors = FALSE)
# Keep top 8 genres for cleaner visuals
top_genres <- df %>%
count(genre, sort = TRUE) %>%
slice_head(n = 8) %>%
pull(genre)
df_genres <- df %>% filter(genre %in% top_genres)
Key Observations:
Release Year Pattern: The bulk of games cluster between 2005 and 2012, reflecting the peak of the PS3/Xbox 360/Wii generation - the most prolific era in gaming history.
Critic Score Concentration: Scores are tightly packed between 7 and 8 across all years and genres, suggesting score inflation or standardized review benchmarks across the industry.
Sales Outliers: A small number of titles - mainly Action and Sports games-spike far above 5 million units, highlighting how blockbuster franchises drive disproportionate revenue.
Combined View: The 3D perspective reveals that high sales are not uniformly tied to high scores - release era and genre matter just as much.
Key Observations:
Weak Positive Trend: Higher critic scores generally correspond to higher sales, but the relationship is loose - many high-scoring games sell modestly while some average-scoring titles sell millions.
Action Dominance: Action games appear most frequently across all score ranges and achieve the highest individual sales peaks, confirming their commercial dominance.
Sports Games: Sports titles cluster at moderate scores (7–8) but still achieve strong sales, likely driven by franchise loyalty rather than critical reception.
Score Ceiling Effect: Very few games score above 9, and those that do tend to be iconic titles from major publishers - quality alone does not guarantee strong sales.
df_genres %>%
group_by(genre) %>%
summarise(
Count = n(),
Mean_Sales = round(mean(total_sales), 3),
Median_Sales = round(median(total_sales), 3),
SD_Sales = round(sd(total_sales), 3),
Mean_Score = round(mean(critic_score), 2)
) %>%
arrange(desc(Mean_Sales))
## # A tibble: 8 × 6 ## genre Count Mean_Sales Median_Sales SD_Sales Mean_Score ## <chr> <int> <dbl> <dbl> <dbl> <dbl> ## 1 Shooter 1480 0.673 0.18 1.57 7.27 ## 2 Sports 2581 0.46 0.2 0.814 7.33 ## 3 Action 2825 0.398 0.14 1.03 7.14 ## 4 Racing 1422 0.368 0.14 0.634 7.26 ## 5 Role-Playing 1483 0.287 0.11 0.624 7.33 ## 6 Misc 1988 0.28 0.11 0.546 7.24 ## 7 Simulation 1116 0.269 0.08 0.555 7.23 ## 8 Adventure 1888 0.172 0.04 0.387 7.26
Key Findings:
Action vs. Sports: Action games lead in total count (2,825 titles) while Sports games have a higher mean sales per title, suggesting Sports releases are fewer but more commercially reliable on average.
Shooter Sales Spike: Despite fewer titles, Shooter games show the highest mean and standard deviation in sales - driven by blockbuster franchises like Call of Duty that skew the average dramatically upward.
Score Consistency: Critic score means are nearly identical across all genres (ranging from 7.1 to 7.4), indicating that reviewers do not systematically favor one genre over another.
Platform and Racing: Racing and Platform genres show moderate sales with low variance, reflecting a stable but niche audience that supports these genres consistently.
model <- lm(total_sales ~ critic_score, data = df) summary(model)
## ## Call: ## lm(formula = total_sales ~ critic_score, data = df) ## ## Residuals: ## Min 1Q Median 3Q Max ## -0.9140 -0.3306 -0.2306 0.0394 19.4665 ## ## Coefficients: ## Estimate Std. Error t value Pr(>|t|) ## (Intercept) -1.352876 0.062091 -21.79 <2e-16 *** ## critic_score 0.234725 0.008519 27.55 <2e-16 *** ## --- ## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 ## ## Residual standard error: 0.7932 on 18830 degrees of freedom ## Multiple R-squared: 0.03875, Adjusted R-squared: 0.0387 ## F-statistic: 759.1 on 1 and 18830 DF, p-value: < 2.2e-16
Comprehensive Analysis:
Statistical Significance: The p-value on critic_score is well below 0.05, confirming that the relationship between critic score and sales is real and not due to random chance.
Coefficient: For every 1-point increase in critic score, total sales increase by approximately 0.235 million units - a modest but real effect.
R² = 0.0388: Critic score explains only 3.9% of the variation in sales. The vast majority of what drives sales - franchise brand, marketing budget, platform install base - is not captured by review scores alone.
Conclusion: Critic scores matter at the margins but are not the primary driver of commercial success. A game can score a 9 and sell modestly, while a 7-rated sports sequel sells millions.
Major Findings:
First: Action and Sports genres dominate total worldwide sales volume, driven by large install bases and recurring franchise releases that guarantee minimum sales floors.
Second: Critic scores are a statistically significant but weak predictor of sales (R² ≈ 0.039), confirming that marketing, platform, and brand loyalty outweigh review performance in commercial outcomes.
Third: The 2007–2010 window was the peak era of game releases, coinciding with the simultaneous maturity of the PS3, Xbox 360, and Wii platforms - the highest-competition period in gaming history.
Fourth: Score inflation is evident - the interquartile range for critic scores clusters tightly around 7.3 regardless of genre, suggesting industry-wide convergence in review standards over time.