It depends on your use case. If you go with a positively skewed distribution, then most values are clustered around the left tail. If you go with a negatively skewed distribution, then it is the opposite one. So if possible share your use case and data set as well. So that we can help you with the steps.