One of the projects that I'm excited to be tackling at work is the development of standard metrics to help guide our actions, a dashboard for Hearth Connection's operations. We'll be working with Michael Quinn Patton in April, a local evaluation guru with a national reputation and someone I hold in high esteem. Between now and then, we're preparing our best thinking about the components of the dashboard, in the hopes that our session with MQP is as thought-provoking and useful as it can be.
One aspect of this challenge that's really struck me is the difference in the kind of measures that are useful on a dashboard. Previously, most of my analytical work has been focused on measures that document the tremendous job our partners have done in helping people transition from homelessness to stable housing. Most of these measures we've done a good job at "pegging": engaging people with long histories of homelessness, facilitating rapid movement from homelessness to housing, helping people stay in housing, and while they're at it, increase their ability to function in other areas of their lives.
The challenge of a dashboard is quite different: what measures -- what gauges on the dash -- communicate useful information to help you decide what to do next? A gauge that's always pegged at 100% is just taking up precious space without communicating any useful information.
The correlated challenge is to have measures that are few in number and remain intuitive and understandable. (Even if it contained a ton of information, there's little point in having a Rube Goldberg gauge if it takes 20 minutes to explain what it measures, or you have no idea how any change in behavior will affect the position of the needle.)
I'm excited about the resources we've found that will help us first decide on some of these metrics and then implement the dash. I've long been a fan of Edward Tufte's work on visualizing information, but the great folks at Juice Analytics and the Particletree have helped folks like me take those ideas a step further, by developing and consolidating tools to make implementing clean, clear, information-rich graphics more attainable for the rest of us.
And so have the folks at The Google. They've recently introduced their Chart API, a free, open-source way to embed charts and diagrams into web pages, just by sending them a properly formatted URL. I'm tremendously excited about the potential for this tool to help us quickly imbed elegant visualizations of data into web products, especially Co-Pilot. To give you a taste of the Google Chart API, here's a Venn diagram of the banners I've recently added to this blog, categorized by whether they depict friends and family, travels and/or motorbikes.
And speaking of travel, I recently came across a stunning and evocative work by artist Seyed Alavi, an aerial rendering of the Sacramento River, woven into a carpet at the Sacramento International Airport. Great concept, and, from the looks of it, an equally impressive implementation. Enough to even make me wish for a layover in Sacramento.
Tuesday, March 11, 2008
Keeping track
Posted by Eric at 8:13 AM
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment