Today I had the pleasure of listening to a talk by Dr. Rong Yan, a research Scientist at Facebook.  Dr. Yan was speaking to IST students in the Cybertorium regarding some of Facebook’s research undertakings.  The talk was enlightening and also peppered with some interesting statistics.  The overall research question for Facebook:

“Can we understand users better and help them to share more content and connect with more people and friends?”

This question led to several sub-questions, three of which Dr. Yan focused on today.

1. What are people talking about in the newsfeed? Facebook users spend 20 billion minutes logged into Facebook each day (I wonder how much of that is idle time?), the majority of this time is spent on the newsfeed page.  Facebook users make 1.5 billion newsfeed updates per day.  Similar to twitter research, Facebook has several initiatives surrounding the analysis of newsfeed data.  One is the General Happiness Index, a Facebook application that takes the text from a newsfeed entry, and creates a happiness index based on positive and negative language use. 
2. What types of advertisements are most useful to users?  Not something I’m terribly interested in, but ads bring Facebook over $1 billion in revenue a year so they are very interested. 
3. Can we suggest face tags for each uploaded photo?  Facebook receives 2.5 billion image uploads each month, but 75% of the images have no friends tagged in them.  Facebook currently is testing an algorithm to do this, and it is suggesting tags successfully 99% of the time.

Remember, all of this data is being used to better understand users, to help them share content and make connections. Why don’t we abstract this to a University?  We could frame a similar research question here at PSU, something like:

 “Can we understand students better and help them to identify and grasp new content, also helping them to connect with faculty and other students with similar interests?”

Penn State certainly has the data, right?  We could easily look at trends in registration data, assisting students to find others with similar interests and create an environment where these students can come together to form study groups, clubs and so on to further their educational experience.  We have all sorts of log data from computers, that illustrate what types of web pages and PDFs students are searching for in our labs and dormitories.  Couldn’t this data be used to help inform class content for faculty?  What about ANGEL?  David Wiley spoke at the February Educause Conference about the mostly-wasted data under the hood of our CMS/LMS systems.   Why can’t we use the connections forged in these systems (emails between student and professor, message board postings, team configurations and all other forms of human-to-human interaction) to help give our students the best chance of success?  Many of these data points will also help drive content decisions, allowing us to get a better understanding of what content works in a course, and what content might need some polishing.

I do understand the massive policy hurdles involved with something like this, and how difficult it would be to build and implement systems that could assist in such a momentous undertaking.  But it’s still worth thinking about how we can leverage all our institutional data to help create a better educational experience for our students.