From MediaPost.com:
Thomson Reuters has partnered with Harvard University’s Berkman Center for Internet & Society to launch the Media Cloud, a set of research tools for tracking online media coverage.
The joint effort is meant to bring some clarity to the vast tangle of news and information on the Web, attempting to answer questions about what types of stories are covered by which media sources, where stories begin, and how blogosphere coverage compares to that of the mainstream media.
The open-source project will debut in the second quarter of 2009.
On the Media Cloud site, the team asked the question, “What Are Your Research Ideas?“.
I’m interested to see the full capabilities of this system when it launches later this year. I’m sure it will reveal some interesting truths about the nature of news network effects. It might also assist in media research projects such as those on media bias. But one question came to mind that I couldn’t shake, so I left this comment on the post:
I’m wondering if Media Cloud will be able to analyze not only information in and around a news story, but statistics on the people who READ a story, pass it on, etc. It seems that in today’s world of immediate global communication, these figures are necessary to draw important conclusions about the relative influence of different news outlets. Do you have any ideas for obtaining or estimating the number of people who read about a given topic for each source covering it? Data such as unique visitors to a page and "send to a friend" emails is typically confidential.
A database of news stories, content analysis, trackbacks, citations, and more, can tell a lot about the distribution and nature of news coverage. But it doesn’t answer the question, “What news source is most influential?“
It seems to me more information is required to answer this question. For example, in studies of the influence of Fox News’ conservative bias, researchers measured the change in voting patterns by households who gained access to Fox News versus those households who did not have access to the channel (the control group).
Nonetheless, Media Cloud sounds like an intriguing research project, and I’m happy to hear that the software will be open-sourced and available to the world for study and collaborative improvement.
Chris, the question about influence is an excellent one. It’s not one that the tool is directly built to answer, but I can think of some experiments where we might combine MC information with other metrics out there that help determine influence. For too long, we’ve used link data as our influence metric, in part because it was really all we had – technorati and others did a great job of sharing that data, and we all had fun playing with it. But good traffic info is critical for these analyses, and data sets like Google Ad Planner are extremely promising. Where Media Cloud can be most useful, I think, is in tracing how ideas spread between different branches of media. Once we’re able to build a graph of that ecosystem, questions of readership and influence will be able to be added in as weights on edges of the graph… In other words, you’re absolutely on the right track – we’re going to need your help and the help of other smart people to figure out how to combine our data with other sets to ask and answer the right questions.