Help with this Research: How safe is this street?

To get more people biking in the region and connect area residents with a safe and low-stress bicycle network, governments need to invest more in building safe streets and trails, but we also need new tools to understand the network and set priorities. We are proud to share this guest post from the Urban Computing Lab at the University of Maryland about their research into bicyclist safety. We hope you’ll contribute to the project.

Over the past two decades, cities across the country have experienced a tremendous growth in cycling. As cities expand and improve their bicycle networks, local governments and bicycle associations are looking into ways of making cycling in urban areas safer. However, one obstacle to decreasing the number of bicycle crashes is the lack of information regarding cycling safety at the street level.

With a cycling safety map, we can select our cycling route wisely. Historically, Bicycle Level of Service (BLOS) models have been used to measure street safety. Unfortunately, these models require extensive information about each particular roadway section, which often times is not available. Instead, this project will provide innovative tools to automatically estimate street safety levels from crowdsourced citizens’ complaints as well as to shed some light into the traffic-related reasons behind such safety values.

Our goal is to build a cycling safety map that fits your perception of cycling safety. We assume that such perception is captured by crowdsourced complaints and concerns raised by citizens regarding bicycle and road-related issues. If this assumption holds true, we can use artificial intelligence (AI) techniques to build cycling safety maps with minimum human labor using crowdsourced citizens’ complaints. Our project uses citizens’ complaints extracted from platforms such as 311 or Vision Zero Input Maps. These platforms contain citizen-generated complaints and comments regarding cycling issues including but not limited to, traffic (e.g., speeding, missing road signs), cycling (e.g., street obstructions, lights) or infrastructure (e.g., pavement or curb conditions) at very detailed spatio-temporal scales.

But first, we need to teach our AI techniques about cycling safety levels per road segment, so that the AI techniques can determine how to make good use of the crowdsourced data. And this is where we need help from cyclists like you! Our AI techniques need to know the cycling safety levels to assess how well we can predict them. So, we are asking cyclists to watch and label as many videos as they can. These labels will be used to train our AI techniques and to develop models that will allow decision makers to automatically draw cycling safety maps exclusively using already existing public complaints (e.g, 311); as well as to understand the reasons behind why certain streets might be safer than others.

Ultimately, bicycle associations might also use these insights to support specific street re-designs based on the evidence from the models.  For this project, we are focusing on cycling safety in D.C as a case study, but we hope to expand to other cities in the near future!

Thank you for helping us and happy rating!

Rate Some Videos!

P.S. If you have any questions do not hesitate to contact us at umdcyclingsafety@gmail.com