Figure 2

What are the intrinsic statistical properties of recipes’ popularity distributions? The figure visualizes the CCDF of the empirical popularity distributions of recipes as well as the best fits for several candidate functions. The truncated power law function is the best fit for the whole empirical distribution (cf. panel A). When analyzing only the tail of the empirical distribution (cf. panel B) we can see that the best $x_{min}$ is very high ($x_{min}=20\text{,}784$), which indicates that power law is not a good fit for the whole range of values. Since the remaining tail of the distribution is short, we can not find statistical significant differences between the fits of the power law, lognormal and exponential function.