Millions commute to work every day in cities and interact with colleagues, customers, providers, friends,  and  strangers.   Commuting  facilitates  the  mixing  of  people  from  distant  and  diverse neighbourhoods,  but  whether  this  has  an  imprint  on  social  inclusion  or  instead,  connections remain segregated is less explored. In this project we use geolocated Twitter messages to track the mobility of thousands of users in the 50 largest metropolitan areas of the United States. We combine information on individual commuting with the online social network of users and socio-economic data on their neighborhoods. We use this unique dataset to understand (1) how mobility shapes social relationships of people from different communities, (2) how online social connections, mobility and physical infrastructure connect neighborhoods inside cities, (3) how the online social network of people from lower income neighborhoods differ from the network of people living in rich areas of cities.