Saturday, February 23, 2013

Urban Data Challenge - First Cut

Apologies to my regular readers (both of you, Hi Mom!).  I'm linking this post from other places, so a little background is in order.

I make an open source toolkit for iOS called WhirlyGlobe-Maply.  The WhirlyGlobe part is a 3D interactive globe.  The Maply bit is a flat map based on the same technology.

On to the point.

Urban Data Challenge: The Team


Occasionally I'll do a hackathon, challenge, or whatever you want to call them.  They're kind of cool, but tiring.  It's apparently been long enough to do another one.

This data challenge is kind of cool.  They've got simple transit data sets for three cities: San Francisco, Zurich, and Geneva covering the same week.  The idea is to analyze, display, animate, whatever.  You can build an app, put up a web site, make an animation, whatever.  Very open ended.

So obviously my focus is going to be on an app, but it's no fun to do this stuff alone.

I hooked up with an old colleague for this one, Michael Dougherty.  A man of mystery, Michael has no web presence to speak of.  He has made this awesome, if under advertised iPhone app.  It involves some decent math, so there's some, as you would say, synergy.

First Cut


The first thing to do is parse a data set and get something up on the screen.  That dictates what we do next.  We tossed the data into a sqlite database and started querying.  On the device.  Cause that's what we're doing.  A native app for the iPad.

Monday morning commute.  Feel the surliness.
So yay, it worked!  Well, it didn't at first, but we fixed it and then it worked.  Yay!

The red is passengers getting off at particular stops and the blue is passengers getting on.  We're seeing buses (no trains) here, and not BART or Caltrain, but you can kinda see where they are.  Oh, and this is San Francisco.

Great, so it's working, now it's time to play.  What does the evening commute look like?

Monday evening commute.  Less hungover, but still surly.
The traffic moves around in interesting ways.  Ways that suggest more analysis.  Which is good, since this is just the first cut.  How's the weekend look?

Woo!  Party at Fort Mason.  Which is over.  So go home.
And that looks good.  It basically makes sense, so our errors are more subtle and insidious.  An excellent first start.