In average, one non mobility user visits 3 application categories and one low mobility user visit 6 application categories in the week. For normal users and high mobility user, the average numbers of visited application categories are 9 and 10 respectively. The number of users who visit less than 5 application categories in the week accounts for 86.45%, 45.42%, 8.61% and 1.75% for non mobility, low mobility, normal mobility and high mobility users, respectively.
We now study the impact of the number of cells a user visit on the application categories the user accesses. To this end, we calculate the proportion of flow number of each application category for users with different mobility pattern.
Fig. 18 illustrates the probabilities that users with different mobility access the number of application categories. It shows the relationship between users’ mobility and applications usage. We can see from Fig. 18 that among all the applications, social network is the leading application for users with different mobility span. And there is no clear correlation (and anti-correlation) between mobility and applications category that people access except social network and search engine. Indeed, for those who keep stationary, 40.58% of flows were generated by social network. Yet, for those who have a large mobility span, the percentage accesses for social network increase to 49%. This result is quite different from the result presented in [29], which concluded that email is the most popular application for those who have a large mobility span and social network shows highly intriguing behavior.