An analysis of steam data to see if the number of reviews on a steam game relate to a positive overall rating.

Variables - X = Total number of reviews - Y - Positive Reviews - Using Simple Linear Regression

## Number of games: 27075

## Average positive review rate: 71.4 %

## Median positive review rate: 76 %

## Average number of reviews: 1212

The positive review rate is: \(\displaystyle \text{Positive Review Rate} = {\text{Positive Ratings} \over \text{Total Ratings}} \cdot 100\)

\(\displaystyle {900 \over 900+100}\cdot100 = 90\%\)

The game’s positive review rate is 90%.

\[Y = \beta_0 + \beta_1X + \epsilon\]

where:

• \(Y\) = Positive review rate
• \(X\) = Total reviews
• \(\beta_0\) = Intercept
• \(\beta_1\) = Slope
• \(\epsilon\) = Random error

The fitted equation is:

\[\hat{Y} = b_0 + b_1X\]

regress_data <- g_clean %>%
  filter(tot_reviews < 500000)

# Create regression model
model <- lm(
  positive_rate ~ tot_reviews,
  data = regress_data
)

# Show results
summary(model)

## 
## Call:
## lm(formula = positive_rate ~ tot_reviews, data = regress_data)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -71.734 -12.972   4.466  17.771  28.705 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept) 7.129e+01  1.427e-01  499.65   <2e-16 ***
## tot_reviews 1.509e-04  1.621e-05    9.31   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 23.32 on 27069 degrees of freedom
## Multiple R-squared:  0.003192,   Adjusted R-squared:  0.003155 
## F-statistic: 86.68 on 1 and 27069 DF,  p-value: < 2.2e-16

## `geom_smooth()` using formula = 'y ~ x'