I recently completed a piece on CNNgo.com about the best places to run in Tokyo. These sort of lists tend to make me uncomfortable, and this one would be impossible unless I ran all of the trails personally -- tricky to say the least.
I decided that the best approach would be to survey the Namban running club, a local group of runners well familiar with good runs around town. So I whipped up a few questions and then proceeded to do what I always do when faced with a data-related project. I emailed Chris Amico to see what he'd do. He wisely advised that I trim the fat, because the last thing you want is for your respondents to get bored.
So I then distilled my original, more lengthy survey to the following five criteria, administered using a Google Spreadsheet form:
- Is this route good for long runs (15k+)?
- Is this route good for short runs (5-15k)?
- Is the route scenic?
- Are there showers nearby?
- Are there lockers nearby?
- Optional: Do you have a favorite spot not on the list?
Hardly perfect, but it should provide some decent insight any running trail. Bob Poulson of Namban indicated that they had a list of suggested runs on their website, and after tacking on a few more I had my survey ready to go. Bob was kind enough to circulate it among the runners, and the feedback I got was solid.
But how to present it? Naturally I wanted to map it, but the data wasn't playing well with Fusion Tables or Many Eyes to start (specifically, some of the Japanese addresses were not well received). Ideally I'd use Drupal's GMap and Location modules, but unfortunately CNNgo isn't yet taking advantage of that feature of Drupal.
I was left to do it manually (college flashbacks) on Google Maps proper, dropping a pin for each location on the map. I then took the graphical images from the survey responses and slapped them down in the Google Map pop-up balloon windows. The other key information I wanted to provide was a route map for each spot. Again I could do it manually on Google maps but drawing each trail would be time-consuming, and even worse, it would be ugly. I decided in favor of Nike+ routes, even though I absolutely abhor Nike Running's website. It's a joke. But runners do use it, and they map routes. More importantly, they had mapped all the routes I needed and I quickly gathered them up. I included a Google Map link too, and threw in Wikipedia and YouTube video links as well for good measure.
I considered using the Namban feedback to rank the routes in order (that would be the most 'empirical' way, right?), but to be honest I think most people don't care which is best but rather which route is closest to them -- hence my emphasis on geographic location. The end result is the following map. Sadly, in the end CNNgo decided to pass on embedding it with the story, but I'm happy to publish here. Again, a huge thanks to the Namban running club for taking the time to pass on such awesome feedback. The embed code is below. I've also posted the graphical survey responses below, including some routes that didn't make the CNNgo piece. Scroll down to check 'em out.
As always, feedback welcome.