From routes to reality: Improving biking data with better context

This article was originally published by Mobility Lab.
Adding context to existing datasets can play a significant role in improving the bicycling experience.

That was the takeaway from the most recent Mobility Lab-sponsored Transportation Techies’ Bike Hack Night VI. Programmers and advocates presented and listened to a handful of bike-themed show-and-tells under the impending threat of rain at REI’s space at Wunder Garten, which provided the Meetup group’s first-ever outdoor session.

Bringing directions into real life

Eric Brelsford explained the process of adding bike lanes to Open Street Map, which he describes as the “Wiki Map” of the world. Many apps and services use this platform for their routing tools, meaning accurate data is key to a reliable user experience. New York City, one such provider, asked Brelsford to help update the city’s bike lane data on OSM so that it didn’t have to build its own routing tool from scratch.

Brelsford compared the New York City government’s comprehensive bike-lane data to what already existed on Open Street Map, identifying discrepancies to be addressed within OSM. Because such an undertaking requires a significant time investment, Brelsford explained that he gamified the process, getting contributors to reconcile one point at a time on the maps in their spare time. The end result is a more comprehensive visualization of bike lanes around the city, allowing cyclists to view safe rides to their destinations. While the project examined New York, it could easily be replicated with bike-lane data in other cities.

NYC bike lanes

Taking a more direct approach to visualization, Ryan Abrahamsen presented Terrain360 and the technology that he uses to create a Google Street View-like depiction of off-street trails throughout the country. Terrain360 started with Abrahamsen hoping to encourage more people to explore the outdoors, and figured that showing them the routes they can take would dispel potential fears of the unknown, like on D.C.’s Anacostia Riverwalk trail, for example. His project has since grown into a full-time effort, with organizations such as the Chesapeake Conservancy looking to document their waterways and trails.

With two boats, a hiking pack, and a fat-tire tricycle, Terrain360 has profiled rivers, hiking trails and biking trails, capturing conditions such as topography and ambient weather. His tricycle carries five DSLR cameras with fish-eye lenses that take a picture every 30 feet, ultimately stitching together about 100,000 images to form a high definition, 360-degree view tied to GPS data. By merging technology with the outdoors, Terrain360 is making hidden-away trails more accessible to the interested-but-concerned crowd.

Unifying datasets

James Graham from the District Department of Transportation explained what he calls “DDOT’s New Centerline.” In the District’s Vision Zero efforts, Graham and colleagues noticed that their data sets of street centerlines, the line segments that represent streets within GIS, were not well integrated with other pieces of information. For example, the agency wouldn’t know by looking at certain centerline data if there is also a bike lane on that street – the lane information existed in a separate dataset. This division of information prevented a comprehensive understanding of road safety in D.C.

Graham and his team have worked to pull every bit of lane data into one set. The new centerline “linear reference” standard allows users to see a cross-section of information at any one point in the line: lane widths, bike lanes, parking spots, and more. “You build data in a way that more closely represents reality,” Graham said.

From there, DDOT and civic hackers can compare any given road’s actual conditions with crash data and better determine what the city can do to make it safer for bicyclists.

Kate Rabinowitz, who runs the blog Data Lens D.C., explained how analyzing sensor data of bike ridership on the region’s trails can provide useful insights to understand usage and bike traffic. However, much like DDOT’s centerline problem, only having one set of information doesn’t provide much actionable information for planners. Raw traffic numbers require context to understand what they mean and how to make informed decisions about infrastructure.

BTWD comparison

Rabinowitz compared Bike to Work Day 2016 rider counts to data from previous years.

By providing information about weather and other conditions along the region’s trails, the data starts to suggest answers to why patterns exist or anomalies occur. Regional events can cause spikes in ridership on certain days, while a trail closure can account for days when it appears nobody has biked past the sensor. For example, integrating weather and trail usage information – a “literal heatmap,” Rabinowitz notes – illustrates the relationship between ridership and temperature, a powerful factor in whether people decide to bike or not.

With longer-term analysis, planners can see the effects of other transportation disruptions, like the current SafeTrack work on D.C.’s Metro, and how they influence changes in cycling behavior during and after such disruptions.

Rather than learning from stationary counters, Justin Molineaux of Baas Bikes explained the power of smartphones to provide “low-cost, data-rich bikeshare” while generating a wealth of detailed usage data. Using existing racks and regular bikes with Bluetooth locks in their seatposts, Baas is provides a flexible, low-cost bikeshare network in communities like college campuses that lack the money to invest in large-scale systems.

The data that Baas Bikes collect informs potential demand and cycling patterns that help the company and communities using the system to consider cyclists in future planning. By tracking bike movements, Baas can see where users tend to ride, and where they tend to look for bikes. Baas even determined the distance that potential riders are willing to walk to a bike – two minutes, or .07 miles – based on where they launch the app relative to the closest bike.

This data informs the systems’ expansion and maintenance – in one example, Molineaux explained a noticeable rise in searches for bikes near a particular building outside of the system. The usage in that area alerted his team to a college residence hall that the system didn’t cover, and now does.

Context helps everyone with decision-making, from the people on bikes to the planners optimizing the system for them. Data doesn’t exist in a vacuum, and understanding the conditions for bicyclists is a key way to help more of them feel comfortable and confident.

Also see: WABA and DCFemTech presented their regional bike lane map at the event.

Photo: Eric Brelsford presenting at the Wunder Garten (M.V. Jantzen, Flickr).

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s