Can machine learning yield insights into where best to construct bicycling infrastructure in order to encourage commuters to bicycle to work?

Can machine learning algorithms outperform traditional statistical techniques at predicting which neighborhoods would respond with a particularly strong uptake in commuter cycling in response to the placement of bicycling infrastructure there?

Those who have an interest in these issues -- whether they be civic associations devoted to bicycling, advocates of livable cities, or urban transport planners -- may want to click around this website to find answers to these questions.

Just one note:  This website and my analysis are definitely a work-in-progress. Both my dataset and analysis could use a bit more development, which I will work on going forward.  That said, I believe the value of my website lies not so much in its current content, but rather in its approach: progressively plugging away on different methods of modeling the data, including machine learning approaches, such that we can better predict how commuters will respond to bicycle infrastructure interventions.  I will keep working on the content, too, though.  In the meantime, I hope that you find value in what I have done so far.