An interesting open-source site is Media Cloud. This is a joint project by the MIT Center for Civic Media and the Berkman Klein Center for Internet & Society at Harvard University. It gives access to data which examines how stories arise and are followed in the media.
See these case studies on Understanding ‘Teen Pregnancy’ Frames Using Media Cloud Tools and how the word immigrant is used in context in US reporting of Headstart projects.
Users must register for access to large scale data.