Twitch Leak on October 6, 2021

Data - Pareto

Author

François Geerolf

Info

# **Last data update**: 01 aoû 2026, 21:10. **Last compile**: 15 aoû 2026, 15:30

Sans Regression

Code 
twitch_leak |>
  filter(Rank <= 200) |>
  ggplot() + theme_minimal() + xlab("Gross Earning") + ylab("Rank") + 
  geom_point(aes(x = GrossEarning / 1000000, y = Rank)) +
  scale_x_log10(breaks = 0.001*c(250, 800, 1000, 1500, 2000, 3000, 4000, 6000, 8000),
                labels = dollar_format(accuracy = .1, suffix = "M")) + 
  scale_y_log10(breaks = c(1, 2, 4, 8, 16, 32, 64, 100, 200))

Fit

Regression

Code 
fit1 <- twitch_leak |>
  filter(Rank <= 200) %>%
  lm(log(Rank) ~ log(GrossEarning / 1000000), data = .)

summ(fit1)
Observations 200
Dependent variable log(Rank)
Type OLS linear regression
F(1,198) 29214.72
R² 0.99
Adj. R² 0.99
Est. S.E. t val. p
(Intercept) 4.39 0.01 788.91 0.00
log(GrossEarning/1e+06) -1.79 0.01 -170.92 0.00
Standard errors: OLS

With fit

Code 
twitch_leak |>
  filter(Rank <= 200) |>
  ggplot() + theme_minimal() + xlab("Gross Earning") + ylab("Rank") + 
  geom_point(aes(x = GrossEarning / 1000000, y = Rank)) +
  scale_x_log10(breaks = 0.001*c(250, 800, 1000, 1500, 2000, 3000, 4000, 6000, 8000),
                labels = dollar_format(accuracy = .1, suffix = "M")) + 
  scale_y_log10(breaks = c(1, 2, 4, 8, 16, 32, 64, 100, 200)) +
  geom_function(aes(colour = paste0("Pente (coeff Pareto): ", round(fit1$coefficients[2],3))), fun = function(x) exp(fit1$coefficients[1] + fit1$coefficients[2]*log(x))) +
  theme(legend.position = c(0.8, 0.9),
        legend.title = element_blank())