To further check the hypothesis whether to see the correlation of the count and tip count I tried clustering with a smaller number and using them as sum and average. Thus one can see from both the clusters that the American restaurants have less tip count but the higher number of user count.