The City of Minneapolis has made a strong commitment to alternate modes of transportation over the last decade. Part of this commitment includes documenting pedestrian and bicycle usage data in the city. Check out this page to see the data reports going back to 2007.
The latest report (for 2012) found a 56% increase in bicycle usage (across 30 benchmark locations) and a 22% increase in pedestrian miles (across 23 benchmark locations) between 2007 and 2012.
Minneapolis partners with Transit for Livable Communities to collect the bike/ped data each year in September. Although the data reports are interesting, I found the "how to" report from Minneapolis just as interesting: Non-Motorized Traffic Counts: Operations and Methodology. It describes how they estimate daily bike traffic volumes. They even discuss getting 24 hour bike/ped data from turning movement counts done at 70 signalized intersections in South Minneapolis to use in their benchmarking (we did these counts as part of the large signal retiming project in 2011 – thanks to JoNette Kuhnau for pointing this out to me).
On a similar note, Jim Mearkle sent me a link to bike statistics from Melbourne, Australia. They're using loops (which Minneapolis doesn't fully trust yet) to count bike traffic. According to the website, Melbourne has seen such a huge spike in bicycle commuting this year that the VicRoads engineers thought the counting stations were broken.
I truly believe part of Minneapolis' success is that they are measuring bike and ped traffic. They're building a feedback loop on what works and what doesn't.