Another early morning for the 6:05AM from Bozeman. 4 AM doesn’t feel so bad when the stars are shining and its 30° F outside.
Today I’m on my way to the Science Online ’11 meeting — in fact, I’m posting this in the air between Bozeman and Minneapolis. This is the first year I’ve been able to attend, having been stymied by conflicting advisory board meetings the past two years.
Humbly joining luminaries from science writing and blogging, my motivation for attending is a bit different. I’m most interested in exploring how we can make use of online tools and communities to make the process of science more transparent to other scientists, more accesible to the public, and in general, easier and more efficient.
Publicly-accessible web-based databases have become an essential component of daily research in biomedical sciences. I’m the project manager and lead developer of one such database. We know from user surveys that a vast majority of our users visit the site every day. Most databases — including ours — are referential in nature. You log on, look something up, and log off. But these resources could be so much more than that. We owe it ourselves to look at success cases in other fields to make these websites more interactive and useful.
At the moment, we are currently in the middle of a ground up rewrite of our site. Inspired by the rise of web 2.0 social media and networking, we’re building a number of new tools into the site not commonly found on biological websites.
For example, can we glean biologically meaningful information from the browsing patterns of users? I’ve tried to do this a number of times in the past using log file analysis with no limited success. In our new site, we’ve built a tool that does this in real-time to collect the most popular objects. When correlated with unique users, we can also use this as an Amazon-style suggest feature (“Users interested in this gene were also interested in gene Y”). We’ve extended this concept to a common “favorite this” design pattern to make possible matches even more relevant.
Features like this that revolve around community intelligence pose interesting questions for privacy and transparency. One approach that we are considering is to only tally and only present results to users who have specifically opted in.
Well, we’re descending below 10K feet. Time to post.