However, as it is difficult to compare the distribution of users with absolute numbers, we normalized the data by getting the fraction of the rides by dividing the number of rides per user per day by the total number of rides per user. The normalized distribution provides a clearer understanding of the way customers and subscribers are using the Citi Bikes across the week. From Figure 3, we can clearly tell that the proportion of subscribers using Citi Bike exceeds that of customers on weekdays, but the proportion of customers exceeds that of subscribers on weekends. This reveals that subscribers are more likely to use Citi Bikes on the weekdays than weekends.