The City of San Diego is investing in substantial improvements in biking infrastructure. A long-term effort like this can be difficult to explain to the community in way they find intuitive and meaningful....

The bike2books project provides a simple and meaningful metric of a neighborhood's bike-a-bility: on average how long does it take to bike safely to your nearest library?

• Libraries are often centrally located in their neighborhoods, they are familiar landmarks, and are used by a wide range of the community. If someone feels comfortable biking to their library, they should feel comfortable biking elsewhere in their neighborhood. At least that's the idea.

• bike2books may also be a useful metric in thinking about where to add bicycle infrastructure. The Bike Master Plan focus on both lager between neighborhood connections (which are important), and local 'choke points'. This tool provides one way to estimate those changes impact on a measure of overall bike-a-bility.

• This is not a complete measure of course. It may not be good metric for commuting, for example. That said, a metric like this might be helpful in San Diego developing a bike friendly brand.

This Python module includes the beginning of a tool set for re-sampling based statistical simulations to analyze the bike "bike-a-bility" of each neighborhood in San Diego. The key idea is to develop a good estimate of average travel time to a library, using a data-driven re-sampling approach (i.e. bootstrapping).

• A "trip" starts at random point in a neighborhood.
• The model then estimates the travel time from that point to the closest library, by bike and car.
• Bike trips are penalized for safety (see next section).
• Car trips have a parking penalty (see next section).
• The average bike travel time for N simulated trips is the bike-a-bility score. Smaller is better.
• When the bike-a-bility score falls below the car travel time, biking can be considered a better choice for efficient safe mobility.

• Bike travel time is penalized when the route does not easily include a bike lane or path. This penalty may not reflect the actual travel time, but for biking to become popular it needs to be seen as safe. Providing several measures of bike-a-bility (e.g. one for time, one for safety) can make understanding and quantifying progress difficult. By converting safety (as measure by bike lane availability) into units of time, we can introduce a single bike-a-bility metric. One that lends itself to comparison with other ways of getting around, and is easily understood.

• Bike results are compared to average travel-time by car, penalized by an estimate of how long it might take to park. Parking penalties are based on nearby parking meter usage.

• In the future travel times will be estimated using the Google Maps API. At present that are very crude estimates based on line-of-sight distance.

git clone https://github.com/parenthetical-e/bike2books into a working directory.

On the command line, move into that directory then into the top-level bike2books directory. From there type, pip install -e .

See ipnyb/API_usage.ipynb

• Library locations: http://seshat.datasd.org/sde/library/libraries_datasd.geojson
• Parking meter usage: http://seshat.datasd.org/parking_meters/treas_meters_2016_pole_by_mo_day_datasd.csv
• Bike paths locations: http://seshat.datasd.org/sde/bike_route/bike_routes_datasd.geojson