Sunday, March 20, 2011

Street Bumps

It's been a bumpy road since the last blog entry.  I write this from Santa Fe after the end of the third quarter of classes (term C).  It seems that I can only find time to blog when WPI is on break these days.  Since the last blog entry, Nick and I were in Santa Fe once before, in early January, so I could set up the projects that are about to begin here next week, marking the official first term of operation of the WPI Santa Fe Project Center.
When we got back to Massachusetts a lot more happened to the car, to the castle, to Nick and to me, but I really don't want to continue with the litany of unfortunate events that have incessantly befallen our family since we left Venice back in December. Suffices to say that we're almost back to normal and happy to be away from Mass jinxes. Nick and I drove west once again for the annual pilgrimage to our Mecca and left the bumpy roads behind us.  Maybe it's the thin air, or the sunny vistas, or the elevation, or the people, but Nick and I really like to be back in Santa Fe.  It feels homey here.
Ironically -- and perhaps not so coincidentally -- this past month has also seen the announcement of the debut of StreetBump, our android app for the crowdsourcing of potholes in Boston, which I blogged about before.  The office of New Urban Mechanics of the City of Boston sponsored the app and announced it through a Boston Globe article on February 9, 2011, and the buzz caught on in major international websites (like Popular Science), as well as in very prestigious international technology blogs, such as Wired UK,  Slashdot (the Apple one!), Engadget and Techmeme, all the way to Il Sole 24 Ore, the main Italian business daily (the equivalent of the Wall Street journal in Italy).  Just today, I was interviewed for an article to appear on L'Organe, a francophone magazine in Montreal, Canada.  The list goes on...
Despite all the buzz, though, so far no journalist has caught on to the real innovation.  Everyone has been mesmerized by the mobile app, but the novelty here is not purely technological, but lies primarily in the crowdsourcing of the data, which allows us to not care about identifying the pothole on the fly, since we can statistically (or perhaps bayesianly) rely on the crowd to confirm its presence though repeat hits (or lack thereof) as more users travel the same roads over time.
The post-processing of our Street Bump data is non-trivial so there is going to be an Innocentive competition for the best algorithm to actually weed out the noise from the data and identify road anomalies.  With the upcoming Innocentive challenge, we are thus crowdsourcing the server-side post-processing and I am looking forward to seeing what "the crowd" can come up with. There is a total of $25K prize on this and I suggest that interested readers of this blog consider participating in the challenge.  Unfortunately, being one of the judges, I can't, which is too bad, because I think that Steve Guerin, Josh Thorp and I could probably come up with a good solution using entropy and other complexity techniques.  
In the meantime, the three of us have launched a new site called citizapps.com which may turn into a company once we have enough monetizable apps to justify it.  As part of our investment in the application, we will retain the IP (Intellectual Property) for StreetBump and will also have access to the winning algorithms from the Innocentive competition.  We will, however, release the current code as open-source, so in the end the whole project will be crowdsourced from "soup to nuts".
StreetBump has created quite a stir, but from my perspective it's a natural evolution of my PhD thesis on City Knowledge and all of the various WPI undergraduate projects I have advised in the past 20 years.  In fact, the original pothole mapper was a consequence of work we did in the early 2000's at the Venice Project Center, where I have been bringing WPI students since 1988, when we designed and built a "Moto Ondoso" (boat wake) measuring device for the City of Venice. The initial Pothole Mapper was a rather simple extension of the Moto Ondoso device, when I put on my "other" hat as Director of the WPI Boston Center...  It was a bulky device, with on-board GPS, accelerometers and a microprocessor, and lots of wires, all inside a weatherproof tupperware box.  This home-built Pothole Mapper device took three Major Qualifying Projects to complete, one of which was cited (first) in an MIT paper.  
More importantly, the original PotholeMapper caught the eye of Mayor Menino of Boston who wanted it installed in his SUV.  With the advent of smartphones, the StreetBump app was the logical next step in this decade-long evolution and it made sense that Mayor Menino would support its development, given its pedigree.   In fact, we may be going full-circle, since the gondoliers' guild in Venice has expressed interest in a modified version of StreetBump which will go back to its origins, by recording the moto ondoso that affects the traditional row-boats of Venice...  From wavy water to bumpy roads and back again. Wouldn't that be an appropriate final clincher to this project?   

Let's hope these apps make the waves and bumps settle down a bit.
Nick and I are ready for calmer seas and smoother roads as we settle into our healthier and most serene Santa Fe lifestyle.