Sparklines, Squirrel Cage, Sunnyvale (again)

It's been a busy couple of weeks.

I went to visit my brother and family in the Omaha metro area. Very fun visit and I took a bunch of photos at the Omaha Henry Dorly Zoo (here's a link). The pictures are stored on my local wiki because I thought it'd be an easy way to get good captions for all the photos. You see, my brother has a family pass to the zoo and I think they go there at least once a week in the summer. Hence, my brilliant 10-year-old nephew knows the name of every animal there. Plus, he gets to learn a little web 2.0 lesson in the process. Here's a youtube link to some little videos from the trip. Pay close attention to the ones about the famous Council Bluffs, Iowa Squirrel Cage Jail. Here's a Wikipedia article explaining the Rotary Jail.

Its crazy, but a lot of people (particularly most of my family until now) still have no idea what a wiki is. They know what Wikipedia is, and sometimes they know that anybody can edit it, but that's where it stops.

I must give a special shout-out to Edward Z. Yang (aka Ambush Commander), who wrote this great MediaWiki Extension that allows batch image uploading if you have good access to the server that your wiki is hosted on. SpecialUploadLocal worked like a charm to get my 80 photos into MediaWiki's mechanical stomach.



Within 10 hours of landing back in California, I went to a one-day class in San Francisco taught by Edward Tufte, a preeminent charts and graphs guru and information design expert. Amazing day. Smart guy. Seriously, until the mid-afternoon when I started getting tired, about every 3 minutes I had some "wow, this guy is really smart" thought running through my brain.

One big idea I took home: when presenting information visually, don't waste a single pixel on anything that doesn't convey information. For example, does that arrow really need to be big, green, and have a drop shadow? Probably not. Make it black or gray with a thin line and your audience will focus on the things on either side of the arrow more easily, rather than being distracted by your design. That, and annotate stuff that shows causality—heck, annotate pretty much everything.

His website has very nice information architecture/design forums that are moderated and full of other smart people saying stuff (and probably fanboy disciples, too). His four books are beautiful, and I got nice hardcover editions of each as part of the class. Now that's what I call takeaway. Here's the Wikipedia article about Tufte for a general introduction.

SF 44-57     This is a sparkline of wins and losses for the San Francisco Giants so far this season, generated at hardballtimes.com. Sparkline is a term Tufte coined to describe data dense graphics that can be displayed inline with text. (Wikipedia link) Basically, you've got a little chart or line graph that's a couple of inches long, as tall as regular text, and can convey a lot of meaningful information using potentially thousands of data points. There's a cool PHP package available to start makin' your own at sparkline.org. I'd love to start doing this with my own cycling statistics.

And the other big news—we signed a lease for a place in Sunnyvale and will move there in August. Oh, so much fun to pack everything up and move it all of 4.2 miles. Once we're settled in, I'm hoping to have a free Saturday afternoon to curl up with Beautiful Evidence.

Alaska pics are up; distributed human computation

After about 8 man-hours of labor, I pared the photos down from 1500 to just under 200. Still a lot, but what can I say--Alaska was amazing. It is on my official list of "places to go back to." Wouldn't the Northern Lights be cool in wintertime?

Here they are, in five parts: Alaska07 pics. And here are a few shaky-hand youtube videos. The teaser image here is an ice cave underneath the Mendenhall Glacier outside Juneau.



Something I read in this month's Wired got me thinking of ways I could cut down on all that post-vacation photo labor. Wouldn't it be great if you could, say, put all 1500 of your vacation photos up on Flickr and have the hive mind of that established community tell you which 100 photos best represent the entire collection?

The article is worth reading, an interview with Luis von Ahn. Luis is the inventor of those squiggly/obscured strings of text used to prevent bots from setting up accounts for various online services. These are pretty darn useful for things like keeping communities legit (i.e., preventing spambots from overrunning Yahoo's game sites or craigslist's apartment listings). The device itself is called a CAPTCHA (Completely Automated Public Turing test to tell Humans and Computers Apart), and they're starting to be used for productive side-projects as well, like having unassuming computer users decipher words that OCR systems can't parse.

Other projects von Ahn has worked on are more like games that get bored cube-jockeys and dorm-dwellers to put their time-wasting cycles to work tagging images for use by massive web search tools. There are a lot of applications of this kind of distributed human computation good for broad, gigantic projects, but I'm curious to see if anyone has ideas of how this can be good for me. The crux here, as the article explains, is trying to make the work something fun, that people will play on their own volition, without you having to, like, pay them. Hit me up with great ideas and/or VC cash any time. ;-)